memory_limiter processor
73 complete collector configs on this site use the memory_limiter processor. Each is a whole file, validated with otelcol-contrib v0.161.0.
- Tail a log fileShip an application log file to your backend, and keep your place in it across collector restarts.Logs
- Receive syslog (RFC 5424 and RFC 3164)Accept modern and legacy syslog on one collector, from network gear and hosts that speak nothing else.Logs
- Receive OTLP over gRPC and HTTPThe front door: accept OTLP from SDKs and other collectors, on both transports, for all three signals.Logs · Metrics · Traces
- memory_limiter and batch, in the right orderThe two processors every pipeline needs, so a traffic spike neither kills the collector nor floods the backend.Logs
- Mask PII in logs (e-mail, card numbers, IBAN, tokens, IPs)Rewrite e-mail addresses, card numbers, IBANs, bearer tokens and IPs to placeholders on the collector, before the data leaves your network.Logs
- Redact everything except an attribute allowlistTurn masking around - name the attributes you keep, and the redaction processor removes everything else.Logs
- Pseudonymise user IDs with a salted hashReplace user.id with a salted SHA-256 on logs and spans - same user, same value, so per-user queries still join.Logs · Traces
- Drop sensitive attributes before exportDelete named attributes, and every captured HTTP request header, from logs, spans and metrics before they are exported.Logs · Metrics · Traces
- Enforce required resource attributes (semantic conventions)Refuse telemetry that cannot be attributed - drop anything without service.name, service.namespace and deployment.environment.name.Logs · Metrics · Traces
- Drop health-check logs and spansStop paying to store Kubernetes probe traffic: drop /healthz, /readyz and /livez logs and spans on the collector.Logs · Traces
- Keep only error logsShip only ERROR and FATAL log records from a noisy source, and drop the rest on the collector.Logs
- Filter logs by severity (WARN and above)Drop DEBUG and INFO on the collector, without silently losing the logs that never set a severity.Logs
- Drop all telemetry from a serviceDrop everything a load generator or synthetic probe sends, by service.name, across logs, metrics and traces.Logs · Metrics · Traces
- Sample traces probabilistically (head sampling)Keep a fixed share of traces, decided by trace ID, so every collector keeps the same traces whole.Traces
- Tail sampling: keep every error and slow traceDecide after the trace is complete: keep every failed or slow trace, and only a small share of the healthy ones.Traces
- Reduce metric cardinality (aggregate away per-request attributes)Strip user and request IDs from metric attributes and sum the points that collapse together, so one series per route remains.Metrics
- Drop the collector's own metricsKeep the collector's own otelcol_* metrics out of the application metrics you pay to store.Metrics
- Deduplicate repeated log recordsCollapse a burst of identical errors into one record with a count, instead of storing it a thousand times.Logs
- Parse JSON log lines into attributesTurn each JSON log line into attributes you can filter and group by, keeping the original line as the body.Logs
- Parse nginx access logs into semantic conventionsTurn nginx combined-format access log lines into OpenTelemetry semantic-convention attributes, with typed status codes.Logs
- Parse Apache access logs into semantic conventionsTurn Apache combined-format access log lines into semantic-convention attributes, with typed status codes and sizes.Logs
- Parse key=value log linesTurn logfmt-style key=value lines into attributes, quoted values with spaces included.Logs
- Use the timestamp in the log line, in the right time zoneStamp each record with the time written in the line, not the time the collector read it - in the zone it was written in.Logs
- Map your log levels to OpenTelemetry severitySet severity_number from your level field, including names like warning, err and crit, so severity filters work.Logs
- Add Kubernetes pod, namespace and deployment to telemetryTag every log, metric and span with the namespace, pod, deployment and node it came from.Logs · Metrics · Traces
- Detect host and cloud resource attributesFill in host, environment and cloud resource attributes on the collector - without overwriting what the SDK already set.Logs · Metrics · Traces
- Set resource attributes the sender left outStamp environment and namespace onto telemetry at the collector, without overwriting what an SDK already set.Logs · Metrics · Traces
- Rename old HTTP attributes to current semantic conventionsMove http.method, http.status_code and http.target to the current names, so old and new services land in the same dashboards.Logs · Traces
- Collect host metrics (CPU, memory, disk, network, processes)CPU, memory, disk, filesystem, network, load and per-process metrics from every host, without a separate node agent.Metrics
- Read systemd journal logs (journald)Ship systemd journal entries for the units you choose, at the priority you choose, instead of tailing files that do not exist.Logs
- Scrape Prometheus metrics endpointsScrape any Prometheus /metrics endpoint - node_exporter, an application, an exporter - into OpenTelemetry metrics.Metrics
- Receive StatsD metricsAccept StatsD counters, gauges and timers from applications that already emit them, and send them on as OpenTelemetry metrics.Metrics
- Receive logs from Fluent Bit and Fluentd (forward)Take logs from existing Fluent Bit or Fluentd agents over the forward protocol, without replacing them first.Logs
- Receive plain log lines over TCPA raw TCP port for appliances and scripts that can only write lines to a socket.Logs
- Receive log messages over UDPA UDP port for senders that fire and forget log messages, one datagram per message.Logs
- Check HTTP endpoints (synthetic availability)Probe your HTTP endpoints from the collector and get status class, duration and error metrics per URL.Metrics
- Check TCP ports (databases, caches)Know when a database or cache port stops accepting connections, from the collector next to it.Metrics
- Send logs and traces to ElasticsearchLogs and traces into Elasticsearch, in the OpenTelemetry-native mapping, with the API key from the environment.Logs · Traces
- Send OTLP over gRPC (with headers and TLS)Forward everything over OTLP/gRPC to a backend or another collector, with a tenant header, compression, a queue and retries.Logs · Metrics · Traces
- Send to Grafana Cloud over OTLPLogs, metrics and traces to Grafana Cloud's OTLP gateway, authenticated with your instance ID and token from the environment.Logs · Metrics · Traces
- Send metrics with Prometheus remote writeMetrics into Prometheus, Mimir, Thanos or any remote-write store, with resource attributes as labels you can query.Metrics
- Send logs to Loki over native OTLPLogs into Loki through its native OTLP endpoint - the way that still works in current collectors.Logs
- Send to DatadogLogs, metrics and traces to Datadog, with the API key from the environment and a queue in front of the exporter.Logs · Metrics · Traces
- Send to Splunk over HECLogs and metrics into Splunk through the HTTP Event Collector, with index and sourcetype set and the token from the environment.Logs · Metrics
- Send to Kafka (per-signal topics)Buffer telemetry in Kafka, one topic per signal, OTLP protobuf encoded, zstd compressed.Logs · Metrics · Traces
- Write telemetry to a file (JSON lines, rotated)Write everything to a rotated JSON-lines file, for debugging, air-gapped hand-over or a quick replay.Logs · Metrics · Traces
- Collect Kubernetes pod logs (DaemonSet)Every container's logs from every node, with namespace, pod and container read from the log path - one DaemonSet, no sidecars.Logs
- Collect node, pod and container metrics from the kubeletCPU, memory, filesystem and network for every node, pod and container, read from the kubelet on each node.Metrics
- Collect Kubernetes cluster metrics (one replica)Deployment replicas, pod phases, node conditions, restarts and resource quotas from the Kubernetes API - one collector for the whole cluster.Metrics
- Collect Kubernetes events as logsScheduling failures, image pull errors, OOM kills and probe failures as log records - kept after Kubernetes deletes the events.Logs
- Scrape cAdvisor container metrics from the kubeletThe container_* metrics Kubernetes dashboards expect, scraped from each node's kubelet with the pod's service account token.Metrics
- Scrape pods by prometheus.io annotationsScrape every pod that asks for it with prometheus.io/scrape annotations, each node's collector only its own pods.Metrics
- Collect Windows Event Log (Application, System, Security)Application, System and Security events as structured logs - provider, event id, level, event data - without the rendered text twice over.Logs
- Parse IIS W3C access logs into semantic conventionsEvery IIS request as a structured log: method, path, query, status, sizes, time taken and user agent, with the request's own UTC timestamp.Logs
- Collect Windows performance counters (CPU, memory, disk)CPU busy time, available memory and free space per volume from Windows performance counters, named the way you choose.Metrics
- Store telemetry in ClickHouseLogs, traces and metrics into ClickHouse tables the exporter creates for you, expired by TTL.Logs · Metrics · Traces
- Archive telemetry to S3 (gzip, time-partitioned)Cheap long-term retention: every batch as a gzipped OTLP JSON object under a year/month/day/hour/minute prefix.Logs · Metrics · Traces
- Send to Azure Monitor (Application Insights)Logs, metrics and traces to Application Insights, authenticated by the connection string from the environment.Logs · Metrics · Traces
- Send to Google Cloud (Logging, Monitoring, Trace)Logs, metrics and traces into Cloud Logging, Cloud Monitoring and Cloud Trace, with GCP resource attributes detected.Logs · Metrics · Traces
- Send to Honeycomb over OTLPTraces, logs and metrics into Honeycomb over OTLP, with the team API key from the environment.Logs · Metrics · Traces
- Send to New Relic over OTLPLogs, metrics and traces to New Relic's OTLP endpoint, license key from the environment, EU region.Logs · Metrics · Traces
- Send to Dynatrace over OTLP (delta metrics)Logs, traces and metrics to Dynatrace's OTLP API, with cumulative counters converted to the delta temporality it requires.Logs · Metrics · Traces
- Send logs and traces to OpenSearchLogs and traces into OpenSearch with its own exporter - not the elasticsearch one.Logs · Traces
- nginx access logs to Elasticsearch, parsed into semantic conventionsTail nginx, parse every request into semantic-convention fields with typed status codes, and index it in Elasticsearch.Logs
- Kubernetes pod logs to Loki, enriched with deployment and clusterEvery pod's logs into Loki over OTLP, labelled with namespace, pod, container, deployment, node and cluster.Logs
- Firewall syslog to a SIEM, with PII maskedFirewall and VPN syslog into the SIEM with user names pseudonymised and e-mail addresses masked - IPs kept for correlation.Logs
- Agents to a gateway tier, load-balanced by trace IDAgents route every span of a trace to the same gateway, so tail sampling on the gateways sees complete traces.Traces
- Buffer through a backend outage with a persistent queueKeep what arrives while the backend is down on disk, and deliver it when the backend is back - even if the collector restarts in between.Logs · Traces
- Route logs to different backends by attribute (routing connector)One collector, several backends: the payments team's logs go to their store, everything else to the central one.Logs
- RED metrics from traces (spanmetrics connector) to PrometheusRequest rate, errors and latency per service and endpoint, computed from the spans you already send - no metrics SDK needed.Metrics · Traces
- Count error logs per service as a metric (count connector)Alert on error rates without storing the logs: count them in the collector and keep only the counter.Logs · Metrics
- Prometheus counters to a delta-only backend (cumulativetodelta)Scrape Prometheus counters and convert them to the delta temporality Dynatrace requires - gauges untouched.Metrics
- Windows event logs, reduced: noisy events dropped, rendered text removedWindows events with the structured fields kept, the duplicated rendered text removed and the chattiest event IDs dropped.Logs