ETCD Integration is currently in Beta.
The existing integration of ETCD makes use of HTTP V2 API and Prometheus Endpoint. Below are the details of the implementation
| Datastream Name |
Implementation Method |
| leader |
V2 API |
| metrics |
Prometheus endpoint |
| self |
V2 API |
| store |
V2 API |
Key points
- The current stable version of ETCD is
3.5 . Version 3.6 is in draft state.
- API V2 is not enabled by default in ETCD version 3.5 . It can be enabled by setting the environment variable
- API V2 will be disabled as part of 3.6 . Below are the lines from the 3.6 release plan.
I made an attempt below to link the metrics that were part of the existing ETCD datasets (leader, self, store) and equivalent metrics in the V3 prometheus metrics.
| V2 field |
V3 Metric |
leader.follower.latency.ms |
peer_round_trip_time_seconds |
leader.follower.success_operations , leader.follower.failed_operations |
proposals_committed_total, proposals_applied_total, proposals_pending, proposals_failed_total, has_leader ( All fields are existing in metrics ds) |
self.uptime, self.start_time |
process_start_time_seconds |
self.recv.*, self.send |
etcd_network_peer_received_bytes_total, etcd_network_peer_sent_bytes_total, etcd_network_peer_round_trip_time_seconds_bucket, peer_sent_bytes_total, peer_received_bytes_total, peer_sent_failures_total, peer_received_failures_total, peer_round_trip_time_seconds |
store.gets, store.sets, store.delete, store.update, store.create |
etcd_debugging_store_writes_total, etcd_debugging_store_expires_total, etcd_debugging_store_reads_total, etcd_debugging_store_watchers |
ETCD Integration is currently in Beta.
The existing integration of ETCD makes use of HTTP V2 API and Prometheus Endpoint. Below are the details of the implementation
Key points
3.5. Version3.6is in draft state.I made an attempt below to link the metrics that were part of the existing ETCD datasets (leader, self, store) and equivalent metrics in the V3 prometheus metrics.
leader.follower.latency.msleader.follower.success_operations,leader.follower.failed_operationsself.uptime,self.start_timeself.recv.*,self.sendstore.gets, store.sets, store.delete, store.update, store.create