Monitoring：etcd monitoring
HertzBeat monitors the etcd key-value store by collecting metrics from the Prometheus metrics endpoint that etcd exposes.
etcd 3.4+ is supported (the database size metric etcd_mvcc_db_total_size_in_bytes replaced the old etcd_debugging_* name in 3.4).
PreRequisites
Make sure HertzBeat can reach etcd's metrics endpoint
etcd exposes Prometheus-format metrics on its client port (default 2379) at the /metrics path. Make sure this address is reachable from HertzBeat:
- If etcd only listens on localhost, or client mutual TLS is enabled on the client port, configure a dedicated metrics listener via
--listen-metrics-urls. It serves the metrics and health-check endpoints; if exposed without TLS, restrict it to a trusted network.
- Access
{metrics-host}:{metrics-port}/metrics (the client port 2379 by default) from the HertzBeat host to confirm metrics data can be fetched.
More information see etcd monitoring documentation.
Configuration parameter
| Parameter name | Parameter help description |
|---|
| Target Host | Monitored IPV4, IPV6 or domain name. Note⚠️Without protocol header (eg: https://, http://) |
| Port | Port of the etcd metrics endpoint, default 2379 when using the client listener |
| Query Timeout | HTTP request timeout in milliseconds, default 6000 |
| HTTPS | Whether to use HTTPS to request the metrics endpoint |
| Headers | Optional extra HTTP request headers |
| Auth Type | Optional Basic/Digest auth if the metrics endpoint sits behind an auth proxy |
| Username / Password | Credentials used when Auth Type is set |
Collection Metric
Metric set：etcd_server_has_leader
| Metric name | Metric unit | Metric help description |
|---|
| hasLeader | none | Whether this etcd member has a raft leader (1=yes, 0=no) |
Metric set：etcd_mvcc_db_total_size_in_bytes
| Metric name | Metric unit | Metric help description |
|---|
| dbSize | MB | Total size of the underlying database physically allocated |
Metric set：etcd_server_leader_changes_seen_total
| Metric name | Metric unit | Metric help description |
|---|
| leaderChanges | none | Total number of leader changes observed |
Metric set：process_cpu_seconds_total
| Metric name | Metric unit | Metric help description |
|---|
| cpuSeconds | second | Cumulative user and system CPU time consumed |
Metric set：process_resident_memory_bytes
| Metric name | Metric unit | Metric help description |
|---|
| memory | MB | Resident memory size of the process |