A short orientation to how the syslog forwarder actually works. Read this once and the rest of the docs will make sense without piecing things together from installation, configuration, and annotations all at once.
How it’s structured
This is a forwarder-only product - there’s no app on the receiving end. The destination is your existing SIEM (QRadar, Microsoft Sentinel, or anything else that receives syslog over TCP or UDP).
- Collectord - the agent. Same binary as the Splunk and ElasticSearch products; only the output is different. Runs as a DaemonSet, control-plane nodes included, plus a single Deployment add-on that watches the Kubernetes API for events.
- Syslog - the destination. Collectord opens a TCP or UDP connection to your syslog server and writes each event as one line: the input’s
syslog.formattemplate filled in, followed by a newline. The manifests’ templates are pipe-delimited - the type, the timestamp and the cluster, then the host, namespace, pod, container or file where they apply, and the message last - with no RFC 5424 header.
The product ships in two flavors with parallel directory structure - one for Kubernetes (collectorforkubernetes-syslog) and one for OpenShift (collectorforopenshift-syslog, plus a Red Hat-certified variant). Both deploy into their own namespace so you can run them alongside any other Collectord instance on the same cluster.
What Collectord forwards
Collectord forwards logs and events; it does not forward metrics or Prometheus data through this integration. The syslog product is logs-focused.
The categories you’ll see arrive at your syslog server:
- Logs - container
stdout/stderr, application logs from files inside the container or on a mounted volume, and host logs: the systemd journal and the*.logfiles under/var/log. - Audit logs -
[input.files::audit-logs]is enabled by default and reads the API server’saudit*.logfiles under/var/log/kube-apiserver/on the control-plane nodes (node-role.kubernetes.io/control-plane, the chart’scontrolPlaneNodeSelector; the OpenShift manifest usesnode-role.kubernetes.io/master). OpenShift writes them there out of the box; on Kubernetes, enable audit logging on the API server with--audit-log-pathin that directory. - Events - Kubernetes / OpenShift API events.
Before 26.10, the Kubernetes manifest read no audit logs.
Object watching (Pods, Deployments, etc.) is not enabled by default. You can opt in by adding [input.kubernetes_watch::*] stanzas to the addon ConfigMap if you want object specs to flow over syslog too - keep in mind that syslog isn’t an ideal transport for large structured payloads.
Inputs, outputs, and pipes
Internally, Collectord is a pipeline of three concepts:
- An input discovers data -
[input.files::logs]reads/var/log,[input.kubernetes_events]watches the API server,[input.app_logs]picks up application log files exposed via annotations. - A pipe transforms events on the way through - replace, hash, extract, override, sample, throttle.
- An output ships the result somewhere -
[output.syslog]is the default;devnullexists for dropping data.
Most of annotations is just “configure pipes and the destination output, scoped to one pod or namespace.”
Annotation prefix
This product sets annotationsSubdomain = syslog, so all annotations use the syslog.collectord.io/ prefix instead of plain collectord.io/. This is so you can run a Splunk-output Collectord and a syslog-output Collectord on the same cluster without their annotations colliding - each only reads its own namespace.
If you want an annotation to apply to every Collectord instance regardless of subdomain, use the collectord.collectord.io/ prefix.
Where configuration comes from
The same setting can be specified in several places. The general layering, highest priority first:
- Pod annotations - most specific.
- Workload annotations - annotations on the Deployment / DeploymentConfig / StatefulSet / DaemonSet propagate to its pods.
- Namespace annotations - apply to every pod in the namespace.
- CRD
Configuration- cluster-level defaults applied by selector. - ConfigMap files -
001-general.confand the role-specific files (002-daemonset.conf,004-addon.conf) mounted into the relevant Collectord pods. The lowest-priority layer.
A CRD Configuration can be marked with force: true to override pod, workload, and namespace annotations - see Annotations → Forcing Cluster Level Annotations.
When the same setting is defined at multiple levels and the wrong value is winning, run collectord describe from inside a Collectord pod - see Troubleshooting → Describe.
Auto-discovery vs explicit opt-in
Some things Collectord finds on its own; others require you to point at them:
| Picked up automatically | Requires annotation or config |
|---|---|
Container stdout / stderr | Application log files inside containers |
Host logs (the journal, *.log under /var/log) | Files on a PVC shared by multiple pods |
| Kubernetes / OpenShift events | API objects (Pods, Deployments, ConfigMaps, …) |
API server audit logs (default [input.files::audit-logs]) | Custom field extractions on container logs |
| Audit logging on the Kubernetes API server itself |
If something you expect isn’t reaching your SIEM, the first question is usually: was this auto-discovered, or do I need to point Collectord at it?
How data flows
1Your container ─┐
2 ├─► [input] ──► [pipe] ──► [output.syslog] ──► one line per event ──► Your SIEM
3Your annotation ┘Collectord runs on the nodes, reads from local sources (filesystem, the API server) and writes one line per event over TCP or UDP to your syslog server. Anything beyond that - parsing, indexing, dashboards - is up to your SIEM.
Where to go next
- New cluster: Installation.
- Tuning what gets forwarded for a specific workload: Annotations.
- Something looks wrong: Troubleshooting.