add alerting rule for pulsar components #56
There are no files selected for viewing
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
| Original file line number | Diff line number | Diff line change |
|---|---|---|
| @@ -0,0 +1,282 @@ | ||
| serverFiles: | ||
| alerting_rules.yml: | ||
| groups: | ||
| - name: broker | ||
| rules: | ||
| - alert: BrokerPodNotHealthy | ||
| annotations: | ||
| summary: Kubernetes Pod not healthy. | ||
| expr: kube_pod_status_ready{condition="true", pod=~".*-broker-\\d+"} == 0 | ||
| for: 10m | ||
| labels: | ||
| severity: error | ||
| - alert: BrokerContainerOOMKilled | ||
| annotations: | ||
| summary: Kubernetes container oom killed. | ||
| expr: (kube_pod_container_status_restarts_total - kube_pod_container_status_restarts_total offset 15m >= 1) and ignoring (reason) min_over_time(kube_pod_container_status_last_terminated_reason{reason="OOMKilled", pod =~".*-(broker)-.*"}[15m]) == 1 | ||
| for: 0m | ||
| labels: | ||
| severity: warning | ||
| - alert: BrokerContainerOOMKilled | ||
| annotations: | ||
| summary: Kubernetes container oom killed. | ||
| expr: (kube_pod_container_status_restarts_total - kube_pod_container_status_restarts_total offset 15m >= 2) and ignoring (reason) min_over_time(kube_pod_container_status_last_terminated_reason{reason="OOMKilled", pod =~".*-(broker)-.*"}[15m]) == 1 | ||
| for: 0m | ||
| labels: | ||
| severity: error | ||
| - alert: BrokerHighPublishLatency | ||
| annotations: | ||
| summary: Pulsar Broker P99 publish latency is over 1 second. | ||
| expr: sum(pulsar_broker_publish_latency{quantile="0.99"}) by (pod, cloud_streamnative_io_cluster) / 1000 > 1 | ||
| for: 15m | ||
| labels: | ||
| severity: error | ||
| - alert: BrokerHighPublishLatency | ||
| annotations: | ||
| summary: Pulsar Broker P99 publish latency is over 1 second. | ||
| expr: sum(pulsar_broker_publish_latency{quantile="0.99"}) by (pod, cloud_streamnative_io_cluster) / 1000 > 1 | ||
| for: 5m | ||
| labels: | ||
| severity: warning | ||
| - alert: SizeableCompactedLedger | ||
| annotations: | ||
| summary: There is a topic that has a sizeable compacted ledger with more than 10 million entries (remove entries_count from the topic name in Grafana) | ||
| expr: sum(pulsar_compaction_compacted_entries_count{}) by (pod, cloud_streamnative_io_cluster) > 10000000 | ||
| for: 5m | ||
| labels: | ||
| severity: error | ||
|
Comment on lines
+41
to
+47
There was a problem hiding this comment. we can remove this alert rule, it's a common scenario to be monitored. |
||
| - alert: TooManyReplicationBacklogs | ||
| annotations: | ||
| summary: There is a topic that has too many replication backlogs. | ||
| expr: max(pulsar_replication_backlog{}) by (cluster,topic,namespace,pod) > 50000 | ||
| for: 5m | ||
| labels: | ||
| severity: error | ||
| - alert: BrokerHighTopicLoadPendingRequests | ||
| annotations: | ||
| summary: High number of Topic is pending in Loading requests | ||
| expr: sum(pulsar_broker_topic_load_pending_requests{}) by (pod, cloud_streamnative_io_cluster) > 900 | ||
| for: 10m | ||
| labels: | ||
| severity: error | ||
| - alert: BrokerHighLookUpPendingRequests | ||
| annotations: | ||
| summary: High number of Topic is pending in Lookup requests | ||
| expr: sum(pulsar_broker_lookup_pending_requests{}) by (pod, cloud_streamnative_io_cluster) > 200000 | ||
| for: 10m | ||
| labels: | ||
| severity: error | ||
| - alert: BrokerHighAuthFailures | ||
| annotations: | ||
| summary: Too many authentication failures | ||
| expr: sum by (pod,cluster)(increase(pulsar_authentication_failures_count{}[1m])) > 100 | ||
| for: 5m | ||
| labels: | ||
| severity: warning | ||
| - alert: PulsarTooManyBacklogs | ||
| annotations: | ||
| summary: Too many backlogs | ||
| expr: sum(pulsar_msg_backlog{}) by (pod, cloud_streamnative_io_cluster) > 10000 | ||
| for: 10m | ||
| labels: | ||
| severity: error | ||
| - alert: PulsarTopicLoadLatencyP99 | ||
| annotations: | ||
| summary: Topic Load Latency P99 too high | ||
| expr: sum by(cluster,pod)(topic_load_times{quantile="0.99"}) / 1000 > 30 | ||
| for: 5m | ||
| labels: | ||
| severity: error | ||
| - alert: PulsarTopicLoadFailed | ||
| annotations: | ||
| summary: Topic failed to Load too much times | ||
| expr: sum by(pod, cloud_streamnative_io_cluster)(delta(topic_load_failed_total[30m])) > 30 | ||
| for: 5m | ||
| labels: | ||
| severity: error | ||
| - name: bookie | ||
| rules: | ||
| - alert: BookieContainerOOMKilled | ||
| annotations: | ||
| summary: Kubernetes container oom killed. | ||
| expr: (kube_pod_container_status_restarts_total - kube_pod_container_status_restarts_total offset 15m >= 2) and ignoring (reason) min_over_time(kube_pod_container_status_last_terminated_reason{reason="OOMKilled", pod =~".*-(bookie|bk)-.*"}[15m]) == 1 | ||
| for: 0m | ||
| labels: | ||
| severity: error | ||
| - alert: BookieContainerOOMKilled | ||
| annotations: | ||
| summary: Kubernetes container oom killed. | ||
| expr: (kube_pod_container_status_restarts_total - kube_pod_container_status_restarts_total offset 15m >= 1) and ignoring (reason) min_over_time(kube_pod_container_status_last_terminated_reason{reason="OOMKilled", pod =~".*-(bookie|bk)-.*"}[15m]) == 1 | ||
| for: 0m | ||
| labels: | ||
| severity: warning | ||
| - alert: BookiePodNotHealthy | ||
| annotations: | ||
| summary: Kubernetes Pod not healthy. | ||
| expr: kube_pod_status_ready{condition="true", pod =~".*-(bookie|bk)-.*"} == 0 AND ON (namespace,pod) kube_pod_info{created_by_kind!~"(Job|<none>)"} | ||
| for: 10m | ||
| labels: | ||
| severity: error | ||
| - alert: BookiePodNotHealthy | ||
| annotations: | ||
| summary: Kubernetes Pod not healthy. | ||
| expr: sum by(mon_poolmember,namespace,pod,job) (label_replace(kube_job_labels{},"pod","$1","job_name","(.*)-decommission") == 1) AND ON (namespace,pod) kube_pod_status_ready{condition="true", pod =~".*-(bookie|bk)-.*"} == 0 AND ON (namespace,pod) kube_pod_info{created_by_kind!~"(Job|<none>)"} | ||
| for: 2m | ||
| labels: | ||
| severity: info | ||
|
Comment on lines
+120
to
+126
|
||
| - alert: BookieDecommissionJobCompletion | ||
| annotations: | ||
| summary: Job did not complete in time | ||
| expr: sum (kube_job_spec_completions{job_name=~".*-decommission"} - kube_job_status_succeeded{job="kube-state-metrics", job_name=~".*-decommission"}) by (mon_poolmember,job,job_name,namespace) >0 | ||
| for: 1h | ||
| labels: | ||
| severity: error | ||
| - alert: BookiePVCUsage | ||
| annotations: | ||
| summary: High PVC usage. | ||
| expr: 100 * (1 - (kubelet_volume_stats_available_bytes{persistentvolumeclaim =~".*-(bk|bookie)-.*"} / kubelet_volume_stats_capacity_bytes{persistentvolumeclaim =~".*-(bk|bookie)-.*"})) > 85 | ||
| for: 5m | ||
| labels: | ||
| severity: warning | ||
| - alert: BookieLedgerFillingUp | ||
| annotations: | ||
| summary: PulsarCluster bookie ledger to be fill up. | ||
| expr: predict_linear(pulsarcluster:bookie_ledger_usage:sum[6h], 8 * 3600) > 95 | ||
| for: 0m | ||
| labels: | ||
| severity: warning | ||
| - alert: BookieTotalLedgerUsage | ||
| annotations: | ||
| summary: Total PVC usage of bookie ledger is high. | ||
| expr: sum by (pod, cloud_streamnative_io_cluster)(bookie_ledger_dir__pulsar_data_bookkeeper_ledgers_0_usage) / count by (pod, cloud_streamnative_io_cluster)(bookie_ledger_dir__pulsar_data_bookkeeper_ledgers_0_usage) > 85 | ||
| labels: | ||
| priority: P2 | ||
| severity: error | ||
| - alert: BookieTotalLedgerUsage | ||
| annotations: | ||
| summary: At least 2 bookies ledger space usage should be less than 75%. | ||
| expr: (count by (pod, cloud_streamnative_io_cluster)(bookie_ledger_dir__pulsar_data_bookkeeper_ledgers_0_usage) - count by (pod, cloud_streamnative_io_cluster)(bookie_ledger_dir__pulsar_data_bookkeeper_ledgers_0_usage > 75)) < 2 | ||
| for: 5m | ||
| labels: | ||
| priority: P3 | ||
| severity: error | ||
| - alert: BookieHighEntryAddLatency | ||
| annotations: | ||
| summary: Bookie Entry add P99 latency is over than 500ms. Possible bottleneck in IO or BW of the disk. | ||
| expr: bookkeeper_server_ADD_ENTRY_REQUEST{quantile="0.99", success="true"} > 500 | ||
| for: 5m | ||
| labels: | ||
| priority: P3 | ||
| severity: error | ||
| - alert: BookieHighEntryReadLatency | ||
| annotations: | ||
| summary: Bookie Entry read P99 latency is over than 10s. | ||
| expr: avg_over_time(bookkeeper_server_READ_ENTRY_REQUEST{quantile="0.99", success="true"}[10m]) /1000 > 10 | ||
|
There was a problem hiding this comment. |
||
| for: 5m | ||
| labels: | ||
| priority: P3 | ||
| severity: error | ||
| - alert: BookieInstanceWritable | ||
| annotations: | ||
| summary: Less than 3 Writable Bookie instances. | ||
| expr: count by(pod, cloud_streamnative_io_cluster)(bookie_SERVER_STATUS == 1) < 3 | ||
| for: 2m | ||
| labels: | ||
| priority: P1 | ||
| severity: critical | ||
| - alert: BookieInstanceReadonly | ||
| annotations: | ||
| summary: Bookie instance in readonly status. | ||
| expr: bookie_SERVER_STATUS == 0 | ||
| for: 5m | ||
| labels: | ||
| severity: warning | ||
| - alert: BookieNotEnoughWritableInstance | ||
| annotations: | ||
| summary: Writable bookies count is less than 3. | ||
| expr: (count by (pod, cloud_streamnative_io_cluster)(bookie_SERVER_STATUS) - count by (pod, cloud_streamnative_io_cluster)(bookie_SERVER_STATUS != 1)) < 3 | ||
| for: 5m | ||
| labels: | ||
| priority: P2 | ||
| severity: error | ||
| - name: zookeeper | ||
| rules: | ||
| - alert: ZKContainerOOMKilled | ||
| annotations: | ||
| summary: Kubernetes container oom killed. | ||
| expr: (kube_pod_container_status_restarts_total - kube_pod_container_status_restarts_total offset 15m >= 1) and ignoring (reason) min_over_time(kube_pod_container_status_last_terminated_reason{reason="OOMKilled", pod =~".*-(zookeeper|zk)-.*"}[15m]) == 1 | ||
| for: 0m | ||
| labels: | ||
| severity: critical | ||
| - alert: ZKPodNotHealthy | ||
| annotations: | ||
| summary: Kubernetes Pod not healthy. | ||
| expr: kube_pod_status_ready{condition="true", pod =~".*-(zookeeper|zk)-.*"} == 0 AND ON (pod) kube_pod_info{created_by_kind!~"(Job|<none>)"} | ||
| for: 10m | ||
| labels: | ||
| severity: critical | ||
| - alert: ZKHighWatchers | ||
| annotations: | ||
| summary: Watchers of Zookeeper server is over than 1200k. | ||
| expr: znode_count{} > 1200000 | ||
| for: 5m | ||
| labels: | ||
| severity: warning | ||
| - alert: ZKLeaderLost | ||
| annotations: | ||
| summary: Zookeeper cluster leader lost. | ||
| expr: count by(cloud_streamnative_io_cluster,pod)(leader_uptime) !=1 | ||
| for: 5m | ||
| labels: | ||
| severity: critical | ||
| - alert: ZKDataInconsistent | ||
| annotations: | ||
| summary: znode count inconsistent. | ||
| expr: sum by (cloud_streamnative_io_cluster,po)(increase(diff_count[1m])) > 10 | ||
| for: 2m | ||
| labels: | ||
| severity: error | ||
| - alert: ZKReadLatency | ||
| annotations: | ||
| summary: ZK P99 ReadLatency too high. | ||
| expr: readlatency{quantile="0.99"} > 500 | ||
| for: 5m | ||
| labels: | ||
| severity: warning | ||
| - alert: ZKUpdateLatency | ||
| annotations: | ||
| summary: ZK P99 UpdateLatency too high. | ||
| expr: updatelatency{quantile="0.99"} > 1000 | ||
| for: 5m | ||
| labels: | ||
| severity: warning | ||
| - alert: ZKPVCUsage | ||
| annotations: | ||
| summary: High PVC usage. | ||
| expr: 100 * (1 - (kubelet_volume_stats_available_bytes{persistentvolumeclaim =~".*-(zk|zookeeper)-.*"} / kubelet_volume_stats_capacity_bytes{persistentvolumeclaim =~".*-(zk|zookeeper)-.*"})) > 85 | ||
| for: 5m | ||
| labels: | ||
| severity: error | ||
| - alert: ZKUnrecoverableError | ||
| annotations: | ||
| summary: ZK in an unexpected state related to incident 846 or 849. Needs restart of ZK. Dashboard needs correction | ||
| expr: unrecoverable_error_count > 0 | ||
| for: 5m | ||
| labels: | ||
| severity: critical | ||
| - name: proxy | ||
| rules: | ||
| - alert: ProxyHighActiveConnections | ||
| annotations: | ||
| summary: The proxies are close to reach the maximum amount of connections. | ||
| expr: sum(pulsar_proxy_active_connections{}) by (pod, cloud_streamnative_io_cluster)>60000 | ||
| for: 10m | ||
| labels: | ||
| severity: error | ||
| - alert: ProxyContainerOOMKilled | ||
| annotations: | ||
| summary: Kubernetes container oom killed. | ||
| expr: (kube_pod_container_status_restarts_total - kube_pod_container_status_restarts_total offset 15m >= 2) and ignoring (reason) min_over_time(kube_pod_container_status_last_terminated_reason{reason="OOMKilled", pod =~".*-(proxy)-.*"}[15m]) == 1 | ||
| for: 0m | ||
| labels: | ||
| severity: error | ||
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
we can remove this one which is duplicated with the above one. user side should determine how long the pending time is.