Skip to main content
RunBook Academy

Glossary

Glossary

Terms used across every course. Course-specific terms are tagged with the course they come from.

3

3-2-1 rule🐧 Linux
The backup sizing rule: 3 copies of the data, on 2

A

ACL🐧 Linux
Access Control List. A list of permissions attached to a
ACL (Access Control List)⬢ Proxmox VE
A list of permissions attached to a path in Proxmox.
Acme / Let's Encrypt⬢ Proxmox VE
Automated certificate management; Proxmox VE supports
Acting set◆ Ceph
The set of OSDs currently responsible for a placement
Action plugin⚙ Ansible
Code that runs on the controller to implement or wrap a task.
Active + clean◆ Ceph
The PG's healthy steady state. The cluster has
Active + degraded◆ Ceph
The PG is missing one or more replicas but
Active-Backup Bond⬢ Proxmox VE
Same as active-passive; common Linux bonding terminology.
Active-Passive Bond⬢ Proxmox VE
A bonding mode where one slave is active and the other is on
Ad-hoc command⚙ Ansible
A single module invocation run from the command line
Affinity Rule⬢ Proxmox VE
HA rule that places a resource on specific nodes (positive) or
Agentless⚙ Ansible
Ansible installs no persistent daemon on managed nodes. It connects
Air-gapped🐧 Linux
A backup copy that is physically unreachable from the network -
Alert◉ Observability
A notification that an event of operational interest has
Alert fatigue◉ Observability
The condition in which an operator stops responding to alerts.
Alerting rule◉ Observability
A Prometheus rule of kind alert. A query, a for: duration,
Alertmanager◉ Observability
The Prometheus component that receives alerts from Prometheus
Alloy◉ Observability
Grafana Alloy is the strategic Grafana collector. It replaces
Annotation◉ Observability
A key-value pair on an alert that carries human-facing context:
Anti-Affinity Rule⬢ Proxmox VE
HA rule that keeps resources on different nodes.
Apdex◉ Observability
Application Performance Index. A normalised score from 0 to 1
Apdex-aligned error rate◉ Observability
The fraction of requests that fall outside the Apdex
API Token⬢ Proxmox VE
Long-lived credential for the REST API, scoped to a user. With
APICv / Posted Interrupts⬢ Proxmox VE
Hardware virtualisation features that improve VM
APM◉ Observability
Application Performance Management. The category of commercial
AppArmor🐳 Docker & Containers
A Linux kernel MAC system. Docker integrates with AppArmor
Application-consistent🐧 Linux
A backup taken after the application has flushed and quiesced
Apply◆ Terraform
The execution of the plan. Apply walks the resource graph,
Argument, build🐳 Docker & Containers
A value passed at build time. docker build --build-arg KEY=VALUE
Argument, runtime🐳 Docker & Containers
A value passed to a container at docker run. `docker run -e
ARP🐧 Linux
Address Resolution Protocol. Maps IP addresses to MAC
ASLR🐧 Linux
Address Space Layout Randomization. A security feature
Attribution◉ Observability
The connection between an aggregate metric datapoint and the
Audit log◉ Observability
A log of system actions: who did what, when, and from where.
Auditability⚙ Ansible
The ability to answer, after the fact: which automation version ran,
Auditd🐧 Linux
The Linux Auditing System. Logs security-relevant events
Authentication (cephx)◆ Ceph
Ceph's default mutual authentication
Auto-scaling🐳 Docker & Containers
The horizontal scaling of containers based on load. Docker

B

Backend◆ Terraform
The storage location for the state. Local backends store the
Backfill◆ Ceph
The operation that populates a newly added OSD (or an
Backpressure◉ Observability
The mechanism by which a downstream consumer signals an
Balancing◆ Ceph
The ongoing process by which Ceph keeps PG counts
Base image🐳 Docker & Containers
A starting image for a Dockerfile's FROM instruction.
Base time series◉ Observability
A time series with no labels, or only labels with single
Bash⬢ Proxmox VE
Default shell on Proxmox nodes. CLI tools (qm, pct, etc.) are bash-friendly.
Bastion / jump host⚙ Ansible
An intermediate host through which the controller reaches managed nodes,
BBU (Battery Backup Unit)⬢ Proxmox VE
A battery or supercapacitor on a RAID controller that
Beacon◉ Observability
A small JavaScript snippet that sends telemetry from a browser
Bind mount🐳 Docker & Containers
A mount that binds a host file or directory into a container.
BIOS⬢ Proxmox VE
Legacy VM firmware; modern VMs use UEFI/OVMF.
Black-box monitoring◉ Observability
Telemetry produced by a probe that simulates an external user
Blast radius⚙ Ansible
The number and criticality of systems a single mistake can affect.
Blast radius◆ Terraform
The amount and criticality of infrastructure potentially
Block⚙ Ansible
block / rescue / always: grouped tasks with error handling, roughly
BlueStore◆ Ceph
Ceph's default object store since Reef (18.x),
BMC🐧 Linux
Baseboard Management Controller. A chip on the
Bond⬢ Proxmox VE
Linux link aggregation. bond-mode 802.3ad is LACP.
BorgBackup🐧 Linux
A deduplicating, compressing, encrypting backup program for
Break-glass🐧 Linux
An access path deliberately kept outside the normal
Break-Glass Account⬢ Proxmox VE
Emergency-use account with TFA disabled. Stored in a sealed
Bridge⬢ Proxmox VE
Linux virtual switch. vmbr0 is the typical default.
Bridge network🐳 Docker & Containers
The default Docker network. Containers on the same bridge can
Btrfs🐧 Linux
A modern copy-on-write filesystem with snapshots, subvolumes,
Btrfs⬢ Proxmox VE
A Linux filesystem with snapshots and checksumming. Less mature in Proxmox than ZFS.
Bucket◆ Ceph
A CRUSH container that holds items (leaves) or other
Bucket◉ Observability
A bin in a histogram. The cumulative count of observations
Build context🐳 Docker & Containers
The file or directory passed to docker build. The Dockerfile
BuildKit🐳 Docker & Containers
The default Docker build engine since Docker 23.0. BuildKit

C

CA🐧 Linux
Certificate Authority. The entity that signs digital
Callback plugin⚙ Ansible
Code that consumes run events and produces output: the human-readable
Canary⚙ Ansible
A deliberately small first batch — one host, or a designated group —
Canary🐧 Linux
The first host, or small group of hosts, to receive a change,
Capability🐧 Linux
A discrete slice of root\'s power that can be granted on its
Cardinal, label◉ Observability
See Cardinality, label.
Cardinality◉ Observability
The number of distinct time series produced by a metric. For
Cardinality analyser◉ Observability
A tool or query that breaks down the series count of a metric
Cardinality Audit◉ Observability
A quarterly review where the platform team inspects top
Cardinality budget◉ Observability
A pre-agreed limit on the cardinality a metric, a service, or
Cardinality budget enforcement◉ Observability
The mechanism by which a platform enforces per-tenant or
Cardinality budget review◉ Observability
Quarterly. Per-service. Includes: top 20 high-cardinality
Cardinality estimation◉ Observability
A back-of-envelope calculation: for a metric with N labels of
Cardinality explosion◉ Observability
The specific failure mode where a Prometheus instance runs out
Cardinality incident◉ Observability
The production scenario where a single metric's cardinality
Cardinality incident◉ Observability
A "cardinality incident" is a term the course uses to denote
Cardinality spike◉ Observability
Same as a cardinality incident: a sudden increase in series
Cardinality test◉ Observability
A CI step that rejects changes whose estimated cardinality
Cardinality vs Cardinal◉ Observability
A "cardinal" in this context is sometimes used (informally) to
Cardinality vs Churn◉ Observability
Churn is the rate at which time series are created and
Cardinality vs Hot Spot◉ Observability
A hot spot is a single high-traffic series, often a counter
Cardinality vs Skew◉ Observability
A skew is a non-uniform distribution of values across labels.
Cardinality vs Spans◉ Observability
Tempo stores trace spans. Each span has attributes. High
Cardinality vs Volume◉ Observability
Two distinct concerns. Cardinality is the number of distinct
Cardinality, alert on◉ Observability
A meta-monitor that fires when total series exceeds a budget.
Cardinality, label◉ Observability
A label that, by its value distribution, contributes to series
Cardinality, scheduled◉ Observability
A scheduled increase in cardinality, e.g. a deploy that adds
Cardinality, stream (Loki)◉ Observability
In Loki, "cardinality" refers to the number of unique
Cardinality, tenant◉ Observability
Multi-tenant observability platforms (Mimir, Cortex) have
Cardinality, top◉ Observability
A high-cardinality label or metric. Top-N cardinality is
Cardinality, value◉ Observability
The number of distinct values a single label takes across all
Cardinality, vector◉ Observability
In OTel/vector.dev terms, the cardinality of a metric is its
Cardinality, virtual◉ Observability
A high-cardinality value that is stored in structured metadata
Ceph⬢ Proxmox VE
Distributed object store and filesystem. Proxmox integrates with Ceph as a
Ceph Manager (MGR)⬢ Proxmox VE
Hosts cluster metrics, PG autoscaling, and orchestrator
Ceph Monitor (MON)⬢ Proxmox VE
Maintains cluster membership and consensus. Requires majority
Ceph OSD (Object Storage Daemon)⬢ Proxmox VE
Stores objects on a single disk. The unit of
cephadm◆ Ceph
The supported orchestrator for Ceph. Deploys and
CephFS⬢ Proxmox VE
Ceph's POSIX-compatible filesystem. Built on top of RADOS via MDS daemons.
cephx◆ Ceph
Ceph's default mutual authentication protocol. Entities
cgroup🐳 Docker & Containers
A Linux kernel feature for resource isolation and accounting.
Cgroup🐧 Linux
Control group. A Linux kernel feature for resource
CHAP (Challenge-Handshake Authentication Protocol)⬢ Proxmox VE
Authentication method for
Check mode⚙ Ansible
--check: Ansible predicts changes without making them. Genuinely useful
Churn◉ Observability
The rate at which time series are created and destroyed. A
Churn◉ Observability
Already covered above under Churn. Cross-reference: the
CIB🐧 Linux
Cluster Information Base. Pacemaker\'s XML representation of the
CIDR🐧 Linux
Classless Inter-Domain Routing. The IP addressing scheme
Circuit breaker◉ Observability
A pattern that prevents cascading failure: when a downstream
CLI⬢ Proxmox VE
Command-line interface. Proxmox's qm, pct, pvesh, etc.
Cluster🐧 Linux
A group of hosts that work together as one. Provides high
Cluster⬢ Proxmox VE
A group of Proxmox nodes that share configuration and quorum.
Cluster Filesystem⬢ Proxmox VE
pmxcfs, mounted at /etc/pve/. Replicated across nodes.
Cluster map◆ Ceph
The aggregate state of MON, OSD, MDS, and PG maps.
Cluster network◆ Ceph
A separate network Ceph uses for OSD-to-OSD
Coded pool◆ Ceph
See EC pool (erasure coding).
Cold Standby⬢ Proxmox VE
A backup site that is powered off and spun up on demand.
Collection⚙ Ansible
The distribution unit for modules, roles, plugins and playbooks, namespaced
Collector (OTel)◉ Observability
The OpenTelemetry Collector is a vendor-neutral pipeline for
Colocation🐧 Linux
A Pacemaker constraint that says two resources must run on the
Comparison operator (in PromQL)◉ Observability
>, <, >=, <=, ==, !=. The == operator in PromQL is
Confirmed event◉ Observability
An alert that has been acknowledged and is being worked on.
Connection plugin⚙ Ansible
How a task reaches its target: ssh (default), local, paramiko,
conntrack🐧 Linux
The netfilter connection tracking table. It is what makes a
Consensus◆ Ceph
The MON-set's agreement on the cluster maps. Ceph
Container🐳 Docker & Containers
A running instance of an image. A container is the
Container⬢ Proxmox VE
LXC container. Shares the host kernel; weaker isolation than a VM.
Container runtime🐳 Docker & Containers
The software that creates and runs containers. Docker uses
containerd🐳 Docker & Containers
The Docker daemon's container runtime, extracted as a
Control group🐳 Docker & Containers
See cgroup.
Control node / controller⚙ Ansible
The machine running ansible-playbook. Holds the inventory, the
Convergence⚙ Ansible
Repeated runs moving a host toward the declared state and then leaving it
Copy-on-write🐧 Linux
An allocation strategy where a shared block is duplicated only
Corosync🐧 Linux
The cluster membership and quorum layer for Pacemaker.
Corosync⬢ Proxmox VE
Cluster membership and messaging layer used by Proxmox.
CPU🐧 Linux
Central Processing Unit. The main processor in a host.
CPU shares🐳 Docker & Containers
The relative weight of CPU time across containers. --cpu-shares.
Cpuset🐳 Docker & Containers
A cgroup setting that binds a container to specific CPUs.
Crash-consistent🐧 Linux
A snapshot taken with no application coordination. The
Cron🐧 Linux
The classic Unix scheduler for periodic jobs. Configured
CRUSH◆ Ceph
Controlled Replication Under Scalable Hashing. Ceph's
CRUSH⬢ Proxmox VE
Ceph algorithm that maps data to OSDs based on the cluster topology.
CRUSH map◆ Ceph
The cluster topology: buckets (hosts, racks, ...
CRUSH rule◆ Ceph
A placement rule expression that combines a
Cuttlebook (PG distribution)⬢ Proxmox VE
Ceph Placement Group count for a pool.

D

Daemon◆ Ceph
A running Ceph process. ceph-mon, ceph-mgr,
Daemon🐧 Linux
A background service on Linux. Runs without user
Daemon⬢ Proxmox VE
A long-running background process (pvedaemon, pveproxy, etc.).
Daemon (dockerd)🐳 Docker & Containers
The Docker daemon. Listens on a Unix socket (or TCP with TLS)
Daemon.json🐳 Docker & Containers
The Docker daemon's configuration file. Default
Dashboard◉ Observability
A Grafana presentation of panels that visualise a service's
Dashboard hierarchy◉ Observability
A pattern where dashboards are organised: top-level "fleet
Dashboard variable◉ Observability
A reusable query parameter. $instance, $region, $env are
Dashboard, wallpaper◉ Observability
A dashboard with many panels of green (all happy) that
Data source◉ Observability
A Grafana connection to a backend (Prometheus, Loki, Tempo,
Data source◆ Terraform
A read-only reference to information that Terraform does not
Datacenter⬢ Proxmox VE
Proxmox's top-level logical grouping. Contains clusters, users,
Datastore⬢ Proxmox VE
A PBS storage target, typically backed by a ZFS volume.
DB device◆ Ceph
The BlueStore metadata device. Holds the RocksDB
Dead man's switch◉ Observability
A monitoring system that alerts when the source of telemetry
dentry🐧 Linux
Directory entry. The kernel object that maps a name to an
Dependency graph◆ Terraform
The directed acyclic graph of resources, modules, and their
Detached mode🐳 Docker & Containers
-d / --detach. The container runs in the background.
Device class◆ Ceph
A CRUSH / OSD attribute (typically hdd, ssd,
Devicemapper🐳 Docker & Containers
A block-device storage driver for Docker. The default on
DHCP🐧 Linux
Dynamic Host Configuration Protocol. Assigns IP addresses
DHCP⬢ Proxmox VE
Dynamic Host Configuration Protocol. Containers and VMs can be configured to
Diff mode⚙ Ansible
--diff: shows the textual difference a task would make to a file. The
Disaster Recovery (DR)⬢ Proxmox VE
Recovering at a different site after a primary-site
Distributed Storage⬢ Proxmox VE
Storage replicated across multiple nodes by software
Distributor (Loki)◉ Observability
The Loki component that receives log streams, validates them
DKMS🐧 Linux
Dynamic Kernel Module Support. Allows kernel modules to be
DM🐧 Linux
Device Mapper. The Linux kernel framework for logical
DNS🐧 Linux
Domain Name System. Resolves hostnames to IP addresses.
Docker Compose🐳 Docker & Containers
The tool for defining and running multi-container applications
Docker Engine🐳 Docker & Containers
The umbrella term for the daemon, the CLI, and the runtime.
Docker Hub🐳 Docker & Containers
The default public registry. Production practice: do not pull
Dockerfile🐳 Docker & Containers
The recipe for an image. Production Dockerfiles: minimal base
DPMO◉ Observability
Defects Per Million Opportunities. A Six Sigma metric from
Drain🐧 Linux
Removing a node from service by moving its work elsewhere
DRBD🐧 Linux
Distributed Replicated Block Device. Replicates a block
DRBD⬢ Proxmox VE
Distributed Replicated Block Device. Older replication technology; largely
Drift⚙ Ansible
Divergence of a host's real state from its declared state, caused by manual
Drift◆ Terraform
The discrepancy between what Terraform believes exists (the
Drop-in🐧 Linux
A .conf fragment under /etc/systemd/system/<unit>.d/ that
DS (distributed store)◆ Ceph
See RADOS.
Dynamic inventory⚙ Ansible
Inventory generated at run time from an external source — a cloud API, a

E

eBPF🐧 Linux
Extended Berkeley Packet Filter. A Linux kernel feature
EC◆ Ceph
Erasure coding. A storage scheme where k data chunks are
Edge (CDN / load balancer)◉ Observability
The boundary between the user and the service. Black-box
EFI Disk⬢ Proxmox VE
A small disk holding UEFI firmware variables. Required for OVMF VMs.
Endpoint (Prometheus)◉ Observability
A single (host:port, path) combination that Prometheus scrapes.
Enterprise Repository⬢ Proxmox VE
The Proxmox enterprise APT repository; requires a
Entrypoint🐳 Docker & Containers
The command and arguments that the container runs at start.
Ephemeral🐧 Linux
Short-lived. In cloud, ephemeral storage is lost when the
Epoch◆ Ceph
A monotonic counter incremented on every map change. Used
Erasure Coding (EC)⬢ Proxmox VE
Storage technique that splits data into chunks with parity,
Erasure-coded pool (EC pool)◆ Ceph
A pool where each object's data
etcd🐧 Linux
A distributed key-value store used by Kubernetes for cluster
EVPN (Ethernet VPN)⬢ Proxmox VE
BGP-based Layer 2/Layer 3 VPN technology. Proxmox SDN uses it
Exec🐳 Docker & Containers
docker exec runs a command in a running container. Used for
Execution environment⚙ Ansible
A container image bundling ansible-core, Python dependencies,
Exemplar◉ Observability
A reference to a trace attached to a histogram bucket.
exemplars◉ Observability
a histogram bucket can carry a sample trace ID.
Exporter◉ Observability
A process that exposes telemetry in a format Prometheus
Expose🐳 Docker & Containers
EXPOSE in the Dockerfile declares the port the container
Expr (expression)◉ Observability
A PromQL expression. The result of an expression is a vector
External Metric Server⬢ Proxmox VE
Proxmox's built-in InfluxDB integration.

F

Fact⚙ Ansible
Information Ansible gathers about a managed node — OS, network, hardware —
Failcount🐧 Linux
Pacemaker\'s per-resource, per-node count of failed operations.
Failover⬢ Proxmox VE
The automatic recovery of a service after a host failure. Provided by
Failure domain◆ Ceph
A level of the CRUSH hierarchy that the CRUSH
Fallback metric◉ Observability
A secondary metric used when the primary metric is unavailable.
FastEC⬢ Proxmox VE
Ceph Tentacle's optimised erasure-coding I/O path. Enables partial reads
Federation◉ Observability
A pattern where a top-level Prometheus scrapes one or more
Fencing🐧 Linux
Forcibly removing a node from a cluster (typically by
Fencing⬢ Proxmox VE
Self-fencing via watchdog; prevents split-brain in HA.
FHS🐧 Linux
Filesystem Hierarchy Standard. The standard Linux
FileStore◆ Ceph
The legacy object store replaced by BlueStore.
Filesystem layer🐳 Docker & Containers
A read-only layer in the union filesystem. A Docker image is
Filesystem, union🐳 Docker & Containers
The layered filesystem that combines all image layers plus a
Filtering (Loki)◉ Observability
The act of selecting log streams by label matchers and lines by
Final assessment◉ Observability
A combination of theory quiz and practical assessment that
Firewall (PVE)⬢ Proxmox VE
Proxmox's host and VM firewall, implemented via iptables.
FLOPS query◉ Observability
A PromQL query that produces a stream of floating-point samples.
Flow log◉ Observability
A network-level log of every connection: source / dest IP
Forks⚙ Ansible
forks: how many hosts Ansible addresses in parallel. More is not
FQCN⚙ Ansible
Fully-qualified collection name: ansible.builtin.package rather than
FQDN🐧 Linux
Fully Qualified Domain Name. The complete hostname including
FQDN, telemetry◉ Observability
A fully-qualified domain name in a metric label. Allowed by
FRRouting (FRR)⬢ Proxmox VE
Open-source routing suite used by SDN fabrics.
Full ratio◆ Ceph
mon_osd_full_ratio. When the cluster reaches this

G

Gateway⬢ Proxmox VE
The IP address that routes traffic off the local subnet.
Gauge◉ Observability
A metric type whose value can go up or down: memory used,
GCC🐧 Linux
GNU Compiler Collection. The standard C compiler on Linux.
GFS2🐧 Linux
Global File System 2. Red Hat\'s clustered filesystem.
GlusterFS🐧 Linux
A distributed filesystem that aggregates bricks across
GPU Passthrough⬢ Proxmox VE
Binding a physical GPU to a VM via IOMMU.
Grafana🐧 Linux
The standard observability dashboard. Connects to
Grafana◉ Observability
The open-source visualisation platform. Grafana is the
Grafana Admin◉ Observability
A Grafana user with full permissions. The default admin should
Granularity◉ Observability
The interval between samples. Also called "resolution" or
Graphite◉ Observability
A time-series database popular before Prometheus. Many
Groupadd🐳 Docker & Containers
A Dockerfile instruction that adds a Linux group. Production
GRUB🐧 Linux
GRand Unified Bootloader. The bootloader on most Linux
Guest Agent⬢ Proxmox VE
QEMU Guest Agent (qemu-guest-agent). Improves backup consistency.
GUI⬢ Proxmox VE
Web interface at https://<host>:8006.

H

HA🐧 Linux
High Availability. The discipline of keeping a service
HA (High Availability)⬢ Proxmox VE
Automatic VM failover on host failure. Built on top of the
HA Group (deprecated in PVE 9)⬢ Proxmox VE
Old-style HA placement rule. Replaced by HA Node
HA Node Affinity Rule⬢ Proxmox VE
PVE 9+ HA placement rule, controlling which nodes a
Handler⚙ Ansible
A task that runs only when notified by a changed task, and by default only
HAProxy🐧 Linux
The standard high-performance TCP/HTTP load balancer.
HAProxy⬢ Proxmox VE
Reverse proxy used by pveproxy to terminate web UI/API connections.
Hard Limit⬢ Proxmox VE
A cap on resource consumption (CPU time, memory, etc.).
hardirq🐧 Linux
Hardware interrupt context: the short, non-preemptible handler
Hardware Virtualisation⬢ Proxmox VE
Intel VT-x / AMD-V. Required for KVM.
Hash, log◉ Observability
The application of a hash function to a log message for
HBA🐧 Linux
Host Bus Adapter. A card that connects the host to
HBA (Host Bus Adapter)⬢ Proxmox VE
A non-RAID storage controller. Required for ZFS and Ceph.
HCL🐧 Linux
Hardware Compatibility List. A list of hardware that
HCL◆ Terraform
The HashiCorp Configuration Language. The structured language
Health check◉ Observability
A simple endpoint that reports the service's liveness.
Health probe◉ Observability
A black-box probe that verifies the service is healthy. Health
Healthcheck🐳 Docker & Containers
HEALTHCHECK in the Dockerfile declares a command the daemon
heartbeat◆ Ceph
Bidirectional probe traffic between OSDs, used to
Hierarchical bucket◆ Ceph
A CRUSH bucket that contains other
Histogram◉ Observability
A metric type that tracks the distribution of observations
Host⬢ Proxmox VE
A single Proxmox node.
Host network🐳 Docker & Containers
--network host removes network isolation. The container shares
HTTP API◉ Observability
A network API exposed over HTTP. The Observability course
Hugepages⬢ Proxmox VE
Large memory pages (2 MB or 1 GB). Reduce TLB pressure.

I

IaC◆ Terraform
Infrastructure as Code. The practice of representing
ID (Identity)⬢ Proxmox VE
Authentication realm; Proxmox PAM, AD, LDAP, OIDC.
Idempotency⚙ Ansible
Running the same automation repeatedly converges on the declared state
Idempotent🐧 Linux
An operation that leaves the same result whether it runs once
iDRAC🐧 Linux
Integrated Dell Remote Access Controller. Dell\'s
iLO🐧 Linux
Integrated Lights-Out. HP\'s out-of-band management
iLO / iDRAC / IPMI⬢ Proxmox VE
Out-of-band management interfaces for HPE, Dell, and generic
Image🐳 Docker & Containers
An immutable, layered filesystem plus configuration. An image
Image layer🐳 Docker & Containers
A read-only delta in the image's filesystem. Each Dockerfile
Image registry🐳 Docker & Containers
A server that hosts Docker images. Docker Hub is the public
Immutable🐧 Linux
A copy that cannot be modified or deleted until its retention
Immutable attribute◆ Terraform
An attribute that, when changed, forces the resource to be
Import◆ Terraform
The adoption of a real-world resource into Terraforms state.
Index (Loki)◉ Observability
The in-memory data structure Loki uses to route log queries
Ingress🐧 Linux
The path from external traffic to a service. In K8s, an
initramfs🐧 Linux
The initial RAM filesystem: a small compressed archive the
inode🐧 Linux
The on-disk structure holding a file\'s metadata and the
Inspect🐳 Docker & Containers
docker inspect returns the full configuration of a container,
Instrument◉ Observability
To add code that emits telemetry. An application is
Integration◉ Observability
A connection between two systems. The platform integrates
Integration test◉ Observability
A test that exercises a system component's external contract.
Inventory⚙ Ansible
The set of managed hosts and their grouping. Static files (INI or YAML) or
IOMMU⬢ Proxmox VE
Intel VT-d / AMD-Vi. Required for PCI passthrough.
IPMI🐧 Linux
Intelligent Platform Management Interface. The standard for
IPVS🐧 Linux
IP Virtual Server. The kernel\'s layer-4 load balancer, driven
iSCSI🐧 Linux
Internet Small Computer Systems Interface. SCSI over TCP/IP.
iSCSI⬢ Proxmox VE
Block storage over TCP. Common SAN protocol.
ISO🐧 Linux
International Organization for Standardization. Publishes
ISO⬢ Proxmox VE
Optical disc image. Proxmox can host ISOs and boot VMs from them.
Issue (alert annotation)◉ Observability
A link from an alert to an external issue tracker. Alertmanager
IT (Information Technology)⬢ Proxmox VE
Generic term for IT operations.

J

Job (Prometheus)◉ Observability
A logical grouping of targets that share a scrape configuration.
Join (PromQL)◉ Observability
Combining vectors based on label matchers. `vector_a + on (l1)
Journalctl⬢ Proxmox VE
Linux tool for viewing systemd logs.
Journald🐳 Docker & Containers
The system log service. Docker logs to journald by default on
Journald🐧 Linux
The systemd journal. A structured log system with indexing
JSON log◉ Observability
A log line that is a JSON object. Structured logs. Loki prefers
Jumbo Frames⬢ Proxmox VE
Ethernet frames larger than 1500 bytes (typically 9000). Useful

K

k + m (k=…, m=…)◆ Ceph
The erasure-coding profile. k is the number
kB / KiB🐧 Linux
kB is 1000 bytes (SI); KiB is 1024 bytes (IEC). Linux tools
Kdump🐧 Linux
The Linux kernel crash dump mechanism. Captures memory on
Keepalived🐧 Linux
A daemon that implements VRRP for IP failover and drives IPVS
Kerberos🐧 Linux
A network authentication protocol using tickets. Used by
Kernel🐳 Docker & Containers
The OS kernel. Containers share the host kernel; they do not
Kernel🐧 Linux
The core of the Linux operating system. Manages hardware,
Kernel Same-page Merging (KSM)⬢ Proxmox VE
Linux feature that deduplicates identical memory
Key (in label)◉ Observability
See Label.
keyring◆ Ceph
A Ceph configuration file containing the entity name,
keytab🐧 Linux
A file holding one or more Kerberos principals and their
Kibana🐧 Linux
The visualisation layer for Elasticsearch. Used for log
kprobe🐧 Linux
A kernel dynamic tracing hook that can be attached to almost
Kubernetes🐧 Linux
The container orchestration system. Manages containerised
Kubernetes◉ Observability
The dominant container orchestration platform. The
Kubernetes (K8s)⬢ Proxmox VE
Container orchestration platform. Not directly part of Proxmox
Kubernetes (out of scope)🐳 Docker & Containers
The dominant container orchestrator. Beyond the scope of
KVM🐧 Linux
Kernel-based Virtual Machine. The Linux hypervisor.
KVM⬢ Proxmox VE
Kernel-based Virtual Machine. Linux's hypervisor.

L

Label🐳 Docker & Containers
A key-value pair attached to an image, container, or volume.
Label◉ Observability
A key-value pair on a metric sample. Labels are the dimensions
Label join◉ Observability
A PromQL operation: vector_a + on (l1) vector_b joins two
Label, baggage◉ Observability
In OpenTelemetry terms, a label that propagates across service
Label, dimension◉ Observability
A label that represents a meaningful operational dimension:
Label, high-cardinality◉ Observability
A label whose value distribution is unbounded. user_id,
Label, low-cardinality◉ Observability
A label whose value distribution is bounded. region,
Label, span attribute◉ Observability
A key-value pair on a span. The span attribute is the
LACP🐧 Linux
Link Aggregation Control Protocol. IEEE 802.3ad. Bundles
LACP (Link Aggregation Control Protocol)⬢ Proxmox VE
IEEE 802.3ad; used for LAG bonding.
Langfuse◉ Observability
An open-source LLM observability platform. Out of scope for
Latency◉ Observability
The time between request and response. A latency metric is
Latency⬢ Proxmox VE
Time to complete one operation.
Layer cache🐳 Docker & Containers
The daemon's cache of intermediate layers. Subsequent builds
Layer, writable🐳 Docker & Containers
The top layer of a container's filesystem. Writes to the
LDAP🐧 Linux
Lightweight Directory Access Protocol. The standard for
LDAP (Lightweight Directory Access Protocol)⬢ Proxmox VE
Authentication directory; Proxmox
librados◆ Ceph
The C library that directly speaks the Ceph protocol.
librbd◆ Ceph
The C library for RBD. Used by QEMU/KVM, Proxmox, OpenStack,
Lifecycle◆ Terraform
A meta-argument that controls how a resource is created,
Link (annotation)◉ Observability
See Issue.
Live Migration⬢ Proxmox VE
Move a running VM from one node to another with minimal downtime.
Load (Prometheus)◉ Observability
The state of a target that is being scraped. up{} == 1
Load average🐧 Linux
The three numbers in uptime and /proc/loadavg: an
logrotate🐧 Linux
The standard tool for rotating, compressing and expiring log
Logs (driver)🐳 Docker & Containers
The mechanism that delivers container logs. Default
Loki◉ Observability
The Grafana log aggregation system. Loki indexes stream
Loki tenant◉ Observability
A multi-tenant unit in Loki. Each tenant has its own stream
Lookup plugin⚙ Ansible
Fetches data at template time on the controller: file contents,
LSM🐧 Linux
Linux Security Module. The kernel framework that SELinux,
LUKS🐧 Linux
Linux Unified Key Setup. The standard for disk encryption
LVM🐧 Linux
Logical Volume Manager. The Linux volume management
LVM (Logical Volume Manager)⬢ Proxmox VE
Linux storage abstraction; Proxmox can use it as a
LXC (Linux Containers)⬢ Proxmox VE
Container runtime used by Proxmox for CT.

M

Machine◉ Observability
A target in PromQL — usually a host or container. The instance
MCE🐧 Linux
Machine Check Exception. The CPU\'s way of reporting
MDS (Ceph Metadata Server)⬢ Proxmox VE
Required for CephFS.
MDS (ceph-mds)◆ Ceph
The CephFS metadata server. Holds the inodes
Mean◉ Observability
Arithmetic average. The mean is sensitive to outliers. For
Mean Time Between Failures (MTBF)◉ Observability
A reliability metric: average time between failures. Useful for
Mean Time To Recovery (MTTR)◉ Observability
A reliability metric: average time from a failure occurring to
Median (p50)◉ Observability
The 50th percentile. Half the values are above, half below. The
Memory limit🐳 Docker & Containers
--memory / -m. The container can use no more than this
Memory Overcommit⬢ Proxmox VE
Configuring more VM memory than the host physically has,
Memory swap limit🐳 Docker & Containers
--memory-swap. The total memory + swap the container can
Metadata (Loki structured)◉ Observability
Per-log metadata that is searchable but does not contribute
Metadata (Prometheus)◉ Observability
Non-time-series information attached to a metric: HELP,
Metadata, scrape◉ Observability
Metadata returned by the scrape. Prometheus labels each sample
Metric◉ Observability
A named time series. Prometheus metrics are <metric_name>{<label=value>,...}.
Metric, custom◉ Observability
A metric defined by the application or by an exporter. The
Metric, native◉ Observability
A metric built into the platform or a standard exporter. up,
MGR (ceph-mgr)◆ Ceph
The manager daemon. Hosts pluggable modules
MIB🐧 Linux
Management Information Base. A database used by SNMP for
Migration, alert◉ Observability
The act of changing an alert's definition. Migrations are
Migration, label◉ Observability
The act of changing a label's value set or meaning. env
min_size◆ Ceph
The minimum number of replicas Ceph requires to be
Mixin (Grafana)◉ Observability
A pre-built set of Grafana dashboards and Prometheus alerts for
mm🐧 Linux
Memory management. The Linux kernel subsystem that
Module⚙ Ansible
A unit of work executed on the target: package, service, template.
Module◆ Terraform
A reusable directory of Terraform configuration that can be
Module version◆ Terraform
The version of a module that the consumer is using. Module
MON (ceph-mon)◆ Ceph
The monitor daemon. Maintains the cluster maps
Mount🐳 Docker & Containers
A filesystem path inside the container that is bound to a
msgr / msgr2◆ Ceph
The Ceph network messenger. msgr2 is the secure
mTLS🐧 Linux
Mutual TLS. Both client and server authenticate with
MTTP🐧 Linux
Mean time to patch: the average elapsed time from a security
MTTR🐧 Linux
Mean time to recovery: the average elapsed time from an
MTTR⬢ Proxmox VE
Mean Time To Repair. A measure of how fast you recover.
MTTR, alert◉ Observability
Mean time to repair for alerts. Measured from when the alert
MTU🐧 Linux
Maximum Transmission Unit. The largest payload an interface
MTU⬢ Proxmox VE
Maximum Transmission Unit. Ethernet default 1500; jumbo 9000.
Multi-Queue⬢ Proxmox VE
VM NIC feature for parallel packet processing.
Multi-tenancy◉ Observability
The ability to host multiple customers / teams / environments
Multipath🐧 Linux
Presenting the same SAN LUN over more than one HBA, fabric or

N

Namespace🐧 Linux
The kernel mechanism that gives a process its own view of one
NAS🐧 Linux
Network Attached Storage. File-level storage over the
Native histogram◉ Observability
A Prometheus histogram encoding (Prometheus 2.50+) that
NBD (Network Block Device)⬢ Proxmox VE
Protocol used for VM disk streaming.
Nearfull⬢ Proxmox VE
Ceph OSD utilisation threshold (default 85 %). Triggers backfill pause.
Negative latency◉ Observability
Impossible. Latency >= 0 always. If a metric shows negative
Nested Virtualisation⬢ Proxmox VE
Running a hypervisor inside a VM. Useful for labs.
Network (Docker)🐳 Docker & Containers
A virtual network that containers attach to. The default
Network namespace🐳 Docker & Containers
Each container has its own network namespace. The container's
Network Namespace⬢ Proxmox VE
Linux feature for network isolation. Foundation of LXC.
Network observability◉ Observability
The subset of observability that focuses on network behaviour:
Nexus Switch⬢ Proxmox VE
Cisco data centre switch. Sometimes referenced for fabric designs.
NFS🐧 Linux
Network File System. The standard for shared file
No-Subscription Repository⬢ Proxmox VE
The Proxmox community APT repository; no support
Node-exporter🐳 Docker & Containers
A Prometheus exporter for host metrics. Used to monitor the
Notification◉ Observability
The act of delivering an alert to a receiver: PagerDuty page,
Notification policy◉ Observability
A set of rules that determines which alerts notify which
Notification, group◉ Observability
A group of related alerts notified together. Group by
NSS🐧 Linux
Name Service Switch. The Linux framework for resolving
NTP🐧 Linux
Network Time Protocol. Synchronises clocks across hosts.
NUMA🐧 Linux
Non-Uniform Memory Access. Memory access time depends on
NUMA⬢ Proxmox VE
Non-Uniform Memory Access. Multi-socket memory topology.
NVM Express (NVMe)⬢ Proxmox VE
Low-latency SSD interface.
NVMe🐧 Linux
Non-Volatile Memory Express. A high-performance storage

O

Object◆ Ceph
The unit of storage in RADOS. By default, sized up to
OCFS2⬢ Proxmox VE
Oracle Cluster File System. Used in older Proxmox configurations.
OCI🐳 Docker & Containers
The Open Container Initiative. The standards body for
OCI🐧 Linux
Open Container Initiative. The standard for container
OIDC (OpenID Connect)⬢ Proxmox VE
Federated authentication; Proxmox supports it as a realm.
on the controller⚙ Ansible
file contents,
OOB🐧 Linux
Out-of-Band. Management that works independent of the
Open vSwitch (OVS)⬢ Proxmox VE
Software-defined virtual switch. Optional in Proxmox.
OpenFabric⬢ Proxmox VE
IS-IS-based routing protocol for data centre fabrics.
OS🐧 Linux
Operating System. The software that manages hardware and
OSD (Object Storage Daemon)◆ Ceph
One process per storage device
OSD map◆ Ceph
The cluster map for the OSD set. Each OSD has an
OSI🐧 Linux
Open Systems Interconnection. The 7-layer model for
OSPF⬢ Proxmox VE
Open Shortest Path First; link-state routing protocol.
OTel◉ Observability
OpenTelemetry. The vendor-neutral framework for metrics,
Outage◉ Observability
The state of a service being unavailable. An SLO is a contract
Overlay (filesystem)🐳 Docker & Containers
The recommended storage driver on modern kernels. overlay2
OverlayFS🐧 Linux
A union filesystem. Used by Docker and other container
OVMF⬢ Proxmox VE
Open Virtual Machine Firmware; UEFI for VMs.

P

Pacemaker🐧 Linux
The cluster resource manager for Linux. Decides where
Packer🐧 Linux
A tool for building machine images. The standard for
Page cache🐧 Linux
The kernel\'s cache of file contents in otherwise-unused RAM.
PAM🐧 Linux
Pluggable Authentication Modules. The stack that decides
Panel (Grafana)◉ Observability
A single visualisation in a dashboard. Each panel has a query
Paxos◆ Ceph
The consensus protocol Ceph's MONs use to maintain
PBS (Proxmox Backup Server)⬢ Proxmox VE
Proxmox's backup product; chunked, deduplicated,
PCI Passthrough⬢ Proxmox VE
Binding a physical PCI device to a VM via IOMMU.
PCP🐧 Linux
Performance Co-Pilot. The standard for system performance
Peering◆ Ceph
The state where a set of OSDs agrees on the
PG (placement group)◆ Ceph
A bucket of objects. Every object is
PG (Placement Group)⬢ Proxmox VE
Ceph's data distribution unit.
PG state◆ Ceph
Each PG has a state describing its placement-group-
pg_autoscale_mode◆ Ceph
An MGR module setting that determines
pg_num◆ Ceph
The current number of PGs in a pool. Visible via
PGT⬢ Proxmox VE
Proxmox Global Tracker. Issues with subscriptions.
PID🐧 Linux
Process Identifier. A unique number for each running
PID namespace🐳 Docker & Containers
Each container has its own PID namespace. The container's
Pinned alert◉ Observability
An alert kept firing regardless of the rule's condition. The
PKI🐧 Linux
Public Key Infrastructure. The system of trust based on
Plan◆ Terraform
The output of terraform plan. The plan is a proposal — a
Play⚙ Ansible
A mapping of hosts to tasks. A playbook is a list of plays; keywords like
pmxcfs⬢ Proxmox VE
Proxmox cluster filesystem; FUSE filesystem mounted at /etc/pve/.
Pool◆ Ceph
A logical container of PGs with a common replication /
Pop, dashboard◉ Observability
A dashboard the operator opens to investigate. A pop dashboard
Port mapping🐳 Docker & Containers
-p / --publish. Maps a host port to a container port. `-p
Power Loss Protection (PLP)⬢ Proxmox VE
Enterprise SSD feature that flushes the write cache
Principal🐧 Linux
A Kerberos identity, written name/instance@REALM - a user
Privilege escalation (`become`)⚙ Ansible
Running a task as another user, usually root, usually through sudo.
Privileged Container⬢ Proxmox VE
An LXC container with host-equivalent UIDs (UID 0 = host
Probe◉ Observability
A black-box test. Blackbox_exporter is the Prometheus
Profile◉ Observability
A trace of CPU / memory / IO over time, used to find
Prometheus🐧 Linux
The standard metrics-based monitoring system. Pulls metrics
Prometheus◉ Observability
The open-source monitoring system. Pull-based, with a
PromQL◉ Observability
Prometheus Query Language. Functional, dimensional, time-series.
PromQL, function◉ Observability
rate, increase, irate, histogram_quantile, sum,
PromQL, instant query◉ Observability
A query evaluated at a single time. Returns a vector.
PromQL, operator◉ Observability
Arithmetic (+, -, *, /, %, ^), comparison
PromQL, range query◉ Observability
A query evaluated over a time range. Returns a matrix.
PromQL, vector◉ Observability
A set of time series at a single time. Most PromQL queries
Promtool◉ Observability
The Prometheus command-line utility. promtool check config,
Provider◆ Terraform
A plugin that translates Terraforms resource operations into
Provider alias◆ Terraform
A configuration that allows multiple instances of the same
Public network◆ Ceph
The network Ceph uses for client I/O. Distinct
Pull🐳 Docker & Containers
docker pull fetches an image from a registry. Production
Pull model◉ Observability
The Prometheus ingestion model. Prometheus connects to the
Push🐳 Docker & Containers
docker push uploads an image to a registry. Production
Pushgateway◉ Observability
A Prometheus component for batch jobs and other short-lived
PXE🐧 Linux
Preboot Execution Environment. Boot a host from the
PXE Boot⬢ Proxmox VE
Network boot; useful for provisioning.

Q

qdevice🐧 Linux
corosync-qdevice, a daemon that runs on each cluster node
QDevice⬢ Proxmox VE
A third-party vote holder for Corosync; enables 2-node clusters.
qemu🐳 Docker & Containers
An instruction emulator. The default for --platform
QEMU🐧 Linux
Quick Emulator. The userspace component of KVM. Emulates
QEMU⬢ Proxmox VE
Userspace machine emulator. Proxmox VMs use QEMU on top of KVM.
QEMU Guest Agent⬢ Proxmox VE
A small daemon inside VMs that lets the host coordinate
QPS◉ Observability
Queries per second. A load metric for a service, often a
Quantile◉ Observability
A value below which a given percentage of observations fall.
Query◉ Observability
A PromQL or LogQL expression. A query produces a vector
Quorum🐧 Linux
The minimum number of nodes in a cluster that must agree
Quorum⬢ Proxmox VE
Majority vote for cluster-wide decisions. With N expected votes,

R

RADOS◆ Ceph
Reliable Autonomic Distributed Object Store. The
RADOS⬢ Proxmox VE
Ceph's reliable distributed object store.
RAID🐧 Linux
Redundant Array of Independent Disks. Combines disks for
RAID⬢ Proxmox VE
Redundant Array of Independent Disks. Hardware or software.
RAIDZ⬢ Proxmox VE
ZFS's software RAID; RAIDZ1, RAIDZ2, RAIDZ3 are parity levels.
RAM🐧 Linux
Random Access Memory. The host\'s main memory.
Rate◉ Observability
The change of a counter per unit time. rate(metric[5m])
RAUC🐧 Linux
Robust Auto-Update Controller. Updates embedded Linux
RBAC🐧 Linux
Role-Based Access Control. Permissions based on roles
RBAC◉ Observability
Role-Based Access Control. Grafana has Grafana RBAC.
RBD⬢ Proxmox VE
RADOS Block Device. Ceph block storage exposed as a Linux device.
RBD (RADOS Block Device)◆ Ceph
Block storage over RADOS. Exposed as
RDNS🐧 Linux
Reverse DNS. Resolves IP addresses to hostnames.
Recording rule◉ Observability
A PromQL rule that stores the result of an expression in a new
Recording rule, testing◉ Observability
Use promtool test rules to run a recorded Prometheus unit
Recovery◆ Ceph
The operation that copies object replicas after an
Red (methodology)◉ Observability
A service-level methodology: Rate, Errors, Duration.
Reduce (Prometheus)◉ Observability
A common PromQL pattern: sum without (instance) (...) over a
Reducer (Loki)◉ Observability
The Loki component that takes a set of streams and returns
Registry🐳 Docker & Containers
A server that hosts Docker images. Production practice: use a
Registry (Prometheus)◉ Observability
The list of targets and their labels. Built by service
Remote state◆ Terraform
The state stored in a remote backend (e.g. S3, GCS, Terraform
Replica◆ Ceph
One copy of an object's data. Most replicated pools
Replication◆ Ceph
The default duplication strategy. size=3 means
Resource◆ Terraform
A managed infrastructure object declared in the configuration.
Resource address◆ Terraform
The full path to a resource in a configuration. For a resource
Resource agent🐧 Linux
The script Pacemaker calls to start, stop and monitor a
Restart policy🐳 Docker & Containers
--restart. The container restart behaviour. no,
Restic🐧 Linux
A deduplicating, encrypted backup program with first-class
Retention◉ Observability
How long data is kept. Prometheus: --storage.tsdb.retention.time
Return on Investment, observability◉ Observability
The ratio of operational value (incidents resolved faster,
RGW (RADOS Gateway)◆ Ceph
The S3- and Swift-compatible object
RHEL🐧 Linux
Red Hat Enterprise Linux. The commercial Linux distribution
RocksDB◆ Ceph
The embedded key-value database BlueStore uses for
Rootless mode🐳 Docker & Containers
Docker daemon running without root privileges. Production
Round-trip time◉ Observability
The time for a packet to travel from sender to receiver and
RPKI🐧 Linux
Resource Public Key Infrastructure. Validates BGP routes
RPM🐧 Linux
RPM Package Manager (originally Red Hat Package Manager). Both
RPO🐧 Linux
Recovery Point Objective. The maximum acceptable data loss,
RPO (Recovery Point Objective)⬢ Proxmox VE
Maximum acceptable data loss in time.
RSS🐧 Linux
Resident Set Size. The amount of physical memory a
rsyslog🐳 Docker & Containers
The default syslog implementation on most Linux distributions.
RTO🐧 Linux
Recovery Time Objective. The maximum acceptable downtime,
RTO (Recovery Time Objective)⬢ Proxmox VE
Maximum acceptable downtime.
Rule◉ Observability
In Prometheus, a query evaluated on a schedule. Two kinds:
Rule, evaluation◉ Observability
Each rule has an evaluation_interval (default 1m). The rule
Run🐳 Docker & Containers
docker run creates a container from an image. The most-used
Run queue🐧 Linux
The set of tasks a CPU has ready to run. vmstat 1 reports it
Runbook🐧 Linux
A procedure written to be followed under pressure by someone
Runbook◉ Observability
A step-by-step procedure for handling a specific operational

S

Sample◉ Observability
A single data point: a metric name, label set, value, and
Sampling (trace)◉ Observability
Deciding which traces to keep. Head sampling decides at
SAN🐧 Linux
Storage Area Network. Block-level storage over a dedicated
SAN⬢ Proxmox VE
Storage Area Network. Block storage shared across nodes.
SAR🐧 Linux
System Activity Reporter. Collects and reports system
Saturation🐧 Linux
The degree to which a resource has work queued that it cannot
Saturation◉ Observability
A measurement of how full a resource is. A queue at 90% is
Save / Load🐳 Docker & Containers
docker save and docker load move images via tarball.
SBD🐧 Linux
STONITH Block Device. A shared storage device used to
Schema (Loki)◉ Observability
The on-disk format of stored chunks. Loki's schema is
Schema (Tempo)◉ Observability
The trace block format. Tempo's schema is a versioned columnar
SCP🐧 Linux
Secure Copy. A command-line tool to copy files over SSH.
Scrape◉ Observability
The act of Prometheus pulling metrics from a target. Scrape
Scrape interval◉ Observability
The interval between scrapes. Configurable per scrape config.
Scrape timeout◉ Observability
The maximum time a scrape can take. Default 10s. Slow targets
Scrub◆ Ceph
A background operation that compares object checksums
Scrub⬢ Proxmox VE
ZFS or PBS operation that reads every block and verifies checksums.
SCTP🐧 Linux
Stream Control Transmission Protocol. A reliable,
SDK🐧 Linux
Software Development Kit. Tools for developing software for
SDN⬢ Proxmox VE
Software-Defined Network. Proxmox's overlay networking.
seccomp🐧 Linux
Secure computing mode. A per-process filter on which syscalls
Seccomp🐳 Docker & Containers
A Linux kernel security module. Docker applies a default
Secret🐳 Docker & Containers
A sensitive value (password, API key, certificate) that the
Self-Fencing⬢ Proxmox VE
A node rebooting itself when it cannot reach the cluster quorum.
SELinux🐧 Linux
Security-Enhanced Linux. Mandatory Access Control (MAC) for
Service🐳 Docker & Containers
A Docker Compose abstraction for a long-running container.
Service Discovery◉ Observability
The mechanism by which Prometheus knows which targets to
Service Level Indicator (SLI)◉ Observability
A measured quantity. The course covers SLIs in Part II.
Service Level Objective (SLO)◉ Observability
A target over an SLI. The course covers SLOs in Part II.
SEV (severity)◉ Observability
A level of operational importance. page (page someone) vs
SIEM🐧 Linux
Security Information and Event Management. A system for
Size◆ Ceph
The desired number of replicas Ceph maintains for a PG.
Slab🐧 Linux
The kernel\'s allocator for its own small objects - dentries,
SLO Burn Rate◉ Observability
The rate at which the SLO error budget is being consumed.
SLO, error budget◉ Observability
The amount of unreliability an SLO permits. If the SLO is
SLOG (Separate Log)⬢ Proxmox VE
ZFS intent log device. Speeds up sync writes.
SMB🐧 Linux
Server Message Block. Microsoft\'s file-sharing protocol.
SMS🐧 Linux
Short Message Service. Text messages. Used for alerts in
SNMP🐧 Linux
Simple Network Management Protocol. The standard for
softirq🐧 Linux
Deferred interrupt work: the part of an interrupt handler the
Span◉ Observability
A single unit of work in a distributed trace. Spans have
SPICE⬢ Proxmox VE
Remote display protocol for VMs.
SPIFFE🐧 Linux
Secure Production Identity Framework for Everyone. A
Split brain🐧 Linux
The state where a cluster has partitioned and both halves
SQL🐧 Linux
Structured Query Language. The standard for relational
SRE🐧 Linux
Site Reliability Engineering. The discipline of running
SSD🐧 Linux
Solid State Drive. Storage without moving parts. Higher
SSD⬢ Proxmox VE
Solid-State Drive.
SSH🐧 Linux
Secure Shell. The standard for remote access. Provides
SSH⬢ Proxmox VE
Secure Shell. Proxmox management uses SSH.
SSHD🐧 Linux
The SSH daemon. The server side of SSH.
SSL🐧 Linux
Secure Sockets Layer. The predecessor to TLS. Often used
SSO🐧 Linux
Single Sign-On. Authenticate once, access multiple
SSSD🐧 Linux
System Security Services Daemon. The usual way a Linux host
Stack🐳 Docker & Containers
A group of services defined in a docker-compose.yml or
Stack◉ Observability
The complete observability platform. The course's stack:
Stale◉ Observability
A time series that has not been scraped in the configured
State◆ Terraform
The record of what Terraform believes exists in the real world.
State Lock◆ Terraform
A coordination mechanism that prevents concurrent state
Steal time🐧 Linux
The percentage of a vCPU\'s time the hypervisor gave to some
STONITH🐧 Linux
Shoot The Other Node In The Head. The Pacemaker name for
Storage Replication⬢ Proxmox VE
ZFS-based replication of VM disks to peer nodes.
Strategy plugin⚙ Ansible
Controls task ordering across hosts. linear (default) keeps hosts in step,
Stream◉ Observability
In Loki, a sequence of log lines with the same label set.
Stretch cluster◆ Ceph
A Ceph cluster deployed across two data
Striping◆ Ceph
Distributing an RBD image's data across multiple
Subscribe (Loki)◉ Observability
The act of querying Loki for logs. Subscriptions are pull-based.
Sum (PromQL)◉ Observability
A common aggregation: sum by (label) (metric). Sums up
Summary◉ Observability
A metric type that computes quantiles at the source and
Superblock🐧 Linux
The filesystem\'s own header: size, block count, free counts,
SUSE🐧 Linux
The Linux distribution from SUSE. The original enterprise
SVC🐧 Linux
Service. In K8s, a Service exposes a set of pods.
Swarm🐳 Docker & Containers
Docker's built-in orchestrator. Production practice: use Swarm
systemd🐧 Linux
The init system and service manager on every mainstream Linux

T

Tag🐳 Docker & Containers
A label on an image. image:tag. Production practice: deploy
Tag (Loki label)◉ Observability
The log-side equivalent of a Prometheus label. Loki's
Tail sampling◉ Observability
A sampling strategy that decides at the end of a trace
Tape (LTO)⬢ Proxmox VE
Linear Tape-Open. Long-term backup media.
Target◉ Observability
A scrape endpoint. A target is identified by host:port.
Task (Astro / MDX)◉ Observability
A side-effect-bearing primitive used by components.
TCP🐧 Linux
Transmission Control Protocol. The standard reliable
TCP socket🐳 Docker & Containers
tcp://0.0.0.0:2375 (no TLS, deprecated) or
TDE🐧 Linux
Transparent Data Encryption. Oracle\'s database encryption
Telemetry◆ Ceph
An opt-in MGR module that reports usage statistics
Telemetry🐧 Linux
The collection of metrics and logs for observability.
Tempo◉ Observability
Grafana's trace storage system. The course uses Tempo with
Tenant◉ Observability
A multi-tenant boundary in Loki, Tempo, Mimir, and Grafana.
Tensor◉ Observability
Out of scope for the observability course.
Tentacle◆ Ceph
The Ceph 20.x line, released November 2025. The
Terraform🐧 Linux
HashiCorp\'s infrastructure-as-code tool. Manages cloud
Terraform◆ Terraform
Terraform is a declarative Infrastructure as Code tool that uses
Test Repository⬢ Proxmox VE
Proxmox's pre-release APT repository. NEVER enable in
Test, alert◉ Observability
Synthetic series injected into Prometheus to verify alert
Test, rule◉ Observability
See Recording rule, testing.
TFA / 2FA / MFA⬢ Proxmox VE
Two-Factor / Multi-Factor Authentication. Mandatory for admin
TFTP🐧 Linux
Trivial File Transfer Protocol. A simple file transfer
Thin provisioning🐧 Linux
Allocating storage capacity on write rather than up front, so
Thin Provisioning⬢ Proxmox VE
Allocating storage on demand rather than up-front.
Throughput◉ Observability
The amount of work done per unit time. A service's throughput
Throughput⬢ Proxmox VE
Bytes per second.
Time series◉ Observability
A sequence of (timestamp, value) pairs identified by metric
TIME_WAIT🐧 Linux
The TCP state the side that closes first holds for 2×MSL
TLS🐧 Linux
Transport Layer Security. The standard for encrypted
TLS⬢ Proxmox VE
Transport Layer Security; HTTPS.
Tmpfs🐳 Docker & Containers
A filesystem mount that is backed by host RAM, not disk.
TOFU🐧 Linux
Trust On First Use. The model where you trust the first
Token (auth)◉ Observability
A credential for an API. Grafana service accounts use tokens.
Tombstone◆ Ceph
A marker left on disk after an object or PG is
Topology🐳 Docker & Containers
A constraint on container placement. Production practice:
TPM🐧 Linux
Trusted Platform Module. A chip for secure boot and
Trace◉ Observability
A directed acyclic graph of spans, identified by a trace ID.
Tracepoint🐧 Linux
A static instrumentation point compiled into the kernel with a
TraceQL◉ Observability
The Tempo query language. Similar in style to PromQL /
Tracing◉ Observability
The act of recording traces. OpenTelemetry SDKs instrument
TRIM⬢ Proxmox VE
SSD maintenance command. Required for SSD performance.
TSDB◉ Observability
Time-Series Database. Prometheus's storage engine. The TSDB
TTL🐧 Linux
Time To Live. A field in IP packets that limits the
TTL◉ Observability
Time To Live. A retention or expiry bound. In Loki, a chunk
Two-Factor Authentication (TFA)⬢ Proxmox VE
Same as MFA. Proxmox supports TOTP, WebAuthn,
Type (metric)◉ Observability
Counter, gauge, histogram, summary. The metric type

U

UAT🐧 Linux
User Acceptance Testing. Testing by the end users to
Ubuntu🐧 Linux
The popular Debian-based Linux distribution from
UDP🐧 Linux
User Datagram Protocol. The standard unreliable transport
UEFI🐧 Linux
Unified Extensible Firmware Interface. The modern BIOS
UEFI⬢ Proxmox VE
Unified Extensible Firmware Interface. Modern VM firmware.
UID🐳 Docker & Containers
A user ID. Each container's processes run as a UID; the
UID🐧 Linux
User Identifier. A unique number for each user account.
ulimit🐧 Linux
The shell interface to per-process resource limits (RLIMIT_*):
Ulimits🐳 Docker & Containers
Process resource limits inside the container. --ulimit.
Union filesystem🐳 Docker & Containers
The layered filesystem. A Docker image is a stack of read-
Union mount🐳 Docker & Containers
See union filesystem.
Unprivileged Container⬢ Proxmox VE
An LXC container with UID mapping; safer than privileged.
Unreachable⚙ Ansible
A host the controller could not connect to at all — distinct from a host
Untag (Loki)◉ Observability
The opposite of tag: a label that is removed from a log stream
Up (Prometheus)◉ Observability
A boolean time series indicating whether a target is being
Up set◆ Ceph
The OSDs that should be serving a PG per the cluster
upmap◆ Ceph
An OSD placement map that overrides the standard
Uptime🐧 Linux
The time since the host was last booted. A long uptime
USB🐧 Linux
Universal Serial Bus. The standard for peripheral
USE (methodology)◉ Observability
Utilisation, Saturation, Errors. A resource-level
Use case (observability)◉ Observability
A category of operational question: "is the system up?",
UTC🐧 Linux
Coordinated Universal Time. The standard time reference
UUID⬢ Proxmox VE
Universally Unique Identifier.

V

Vagrant🐧 Linux
HashiCorp\'s tool for building and managing virtual
Vagrant⬢ Proxmox VE
A tool for managing reproducible dev environments. Not directly part of
Validation (config)◉ Observability
The act of checking a config file against the schema. Promtool
Vault⚙ Ansible
Ansible's symmetric encryption for files and strings, so secrets can live
Vault ID⚙ Ansible
A label allowing multiple vault passwords to coexist, typically one per
VCPUs⬢ Proxmox VE
Virtual CPUs assigned to a VM.
VCS🐧 Linux
Version Control System. Examples: git, svn. Git is the
VDO🐧 Linux
Virtual Data Optimizer. A Linux device mapper target for
Vector (PromQL)◉ Observability
A set of time series at a single instant. Most PromQL queries
Vendor◉ Observability
A provider of commercial observability software or service.
Versioning◆ Ceph
A CephX mechanism that signs every protocol
Versioning◉ Observability
Tagging telemetry with version, environment, region. Without
VictoriaMetrics◉ Observability
A Prometheus-compatible time-series database popular for
VIM🐧 Linux
Vi IMproved. The standard text editor on Linux.
VIP🐧 Linux
Virtual IP. An IP address that can move between hosts
VLAN🐧 Linux
Virtual LAN. A layer-2 broadcast domain. Used to separate
VLAN⬢ Proxmox VE
Virtual LAN. 802.1Q tagging.
VLAN-Aware Bridge⬢ Proxmox VE
A Linux bridge that handles VLAN tagging internally.
VM⬢ Proxmox VE
Virtual Machine. KVM/QEMU in Proxmox.
VMID⬢ Proxmox VE
Unique numeric identifier for a VM in the cluster.
VMM🐧 Linux
Virtual Machine Manager. A hypervisor (e.g. KVM, QEMU).
VNC🐧 Linux
Virtual Network Computing. A graphical desktop sharing
VNC⬢ Proxmox VE
Virtual Network Computing. Remote console protocol.
VNet⬢ Proxmox VE
Proxmox SDN's virtual network. A bridge under SDN control.
Volume◆ Ceph
Conceptually distinct from a Ceph pool: a volume in
Volume🐳 Docker & Containers
A persistent storage area managed by Docker. Volumes live
Volume (logs)◉ Observability
Bytes per unit time ingested. Loki's ingest rate is bounded
Volume (metrics)◉ Observability
Samples per unit time scraped. prometheus_tsdb_head_samples
VPP🐧 Linux
Vector Packet Processing. A high-performance network
VRRP🐧 Linux
Virtual Router Redundancy Protocol. The protocol behind
VxLAN🐧 Linux
Virtual Extensible LAN. A layer-2 overlay network over
Vzdump⬢ Proxmox VE
Proxmox's backup tool. Supports snapshot, suspend, stop modes.

W

WAF🐧 Linux
Web Application Firewall. Filters HTTP traffic for known
WAL (write-ahead log)◆ Ceph
A small, sequential-write component of
WAL (Write-Ahead Log)◉ Observability
A persistence pattern: write to log before applying. Used by
Watchdog⬢ Proxmox VE
Hardware or software timer used for self-fencing in HA.
WebAuthn⬢ Proxmox VE
A modern 2FA standard (YubiKey, etc.).
White-box monitoring◉ Observability
Telemetry produced by the system about itself: counters,
Window (PromQL)◉ Observability
A time range. [5m] is a 5-minute window. Windows affect
wipefs🐧 Linux
The tool that erases filesystem, RAID and partition-table
WireGuard⬢ Proxmox VE
Modern VPN protocol. Used by Proxmox SDN for encrypted fabrics.
Witness🐧 Linux
A third party that breaks a tie so an even-sized cluster can
WMI🐧 Linux
Windows Management Instrumentation. A Microsoft framework
Workflow◉ Observability
A sequence of steps an operator follows. Runbooks are
Workspace◆ Terraform
A Terraform feature that allows multiple states in a single
WORM🐧 Linux
Write Once Read Many. Storage that cannot be modified or
Writable layer🐳 Docker & Containers
The topmost layer of a container's filesystem. All writes
WWN⬢ Proxmox VE
World Wide Name. FC identifier.

X

X-axis (graph)◉ Observability
Time. Prometheus graphs are time-series. The x-axis is time
X-pipeline (PromQL)◉ Observability
The right side of a binary operator.
X.509🐧 Linux
The standard format for digital certificates. Used by TLS.
X.509⬢ Proxmox VE
Standard for digital certificates.
X86-64-v3⬢ Proxmox VE
A x86-64 microarchitecture level (AVX2). Recommended CPU model for VMs.
XFS🐧 Linux
A high-performance journaling filesystem. Default on RHEL.
xfs / ext4🐳 Docker & Containers
Filesystem choices for the Docker data root. Production
xlock◆ Ceph
A leadership lock used during recovery coordination.
XML🐧 Linux
Extensible Markup Language. A text-based data format. Used
XZ🐧 Linux
A compression format. Used for Linux kernel images and

Y

Y-axis (graph)◉ Observability
Value. The metric's measurement.
YAML🐧 Linux
YAML Ain\'t Markup Language. A text-based data format. Used
YAML◉ Observability
The data serialisation language used throughout the platform.
YAML⬢ Proxmox VE
Human-readable data format; used for lesson frontmatter.
Yanking the OSD out◆ Ceph
The correct procedure for removing a
YubiKey⬢ Proxmox VE
Hardware token for MFA.

Z

Z-score◉ Observability
A standard score: (x - mean) / stddev. Useful for anomaly
Zabbix🐧 Linux
An open-source monitoring system. Alternative to
Zero (PromQL)◉ Observability
A literal zero. Used in arithmetic to produce baselines:
Zero-day◉ Observability
A new failure mode, previously unseen. Zero-day observability
ZFS🐧 Linux
Zettabyte File System. A combined file system and volume
ZFS⬢ Proxmox VE
A filesystem and volume manager with checksumming, snapshots, and raid
Zombie🐧 Linux
A process that has exited but whose parent has not
Zone🐧 Linux
A logical group in firewalld or BIND. Used to apply
Zvol⬢ Proxmox VE
A ZFS block device. Used as the backing for VM disks on ZFS storage.

`

`--extra-vars` / `-e`⚙ Ansible
Variables supplied on the command line. Wins over almost every other
`--limit`⚙ Ansible
Restricts a run to a subset of the targeted inventory. The single most
`ansible-core`⚙ Ansible
The engine: the CLI tools, the execution model, and the ansible.builtin
`ansible.cfg`⚙ Ansible
The configuration file. Resolved in order: ANSIBLE_CONFIG (environment),
`any_errors_fatal`⚙ Ansible
Play keyword: a failure on any host aborts the play for all hosts.
`changed_when`⚙ Ansible
Overrides a task's change reporting with an expression. A precision
`changed`⚙ Ansible
Ansible's per-task assertion that it altered the system. Not cosmetic:
`count`◆ Terraform
A meta-argument that creates N instances of a resource.
`delegate_to`⚙ Ansible
Runs a task on a host other than the current one, while keeping the
`for_each`◆ Terraform
A meta-argument that creates one instance of a resource per
`ignore_errors`⚙ Ansible
Continues the play despite a failed task. Occasionally correct, usually a
`max_fail_percentage`⚙ Ansible
Aborts a play when the proportion of failed hosts exceeds a threshold. The
`moved` block◆ Terraform
A declarative resource-address migration. The moved block
`no_log`⚙ Ansible
Suppresses a task's arguments and results from output and logs. Necessary
`PLAY RECAP`⚙ Ansible
The per-host summary at the end of a run: ok, changed, unreachable,
`requirements.yml`⚙ Ansible
Declares collection and role dependencies with versions. Pinning here is
`run_once`⚙ Ansible
Runs a task a single time rather than per host. Interacts with serial in
`serial`⚙ Ansible
Splits a play into sequential batches — a count or a percentage. The