EventStore Prometheus exporter

EventStoreDB (https://eventstore.com/eventstoredb/) metrics Prometheus exporter.

Installation

From source

You need to have a Go 1.14+ environment configured.

go get github.com/marcinbudny/eventstore_exporter
cd $GOPATH/src/github.com/marcinbudny/eventstore_exporter 
go build -o eventstore_exporter
./eventstore_exporter \
    --eventstore-url=https://localhost:2113 \
    --eventstore-user=admin \
    --eventstore-password=changeit \
    --cluster-mode=single \
    --insecure-skip-verify

Using Docker

docker run -d -p 9448:9448 \
    -e EVENTSTORE_URL=https://my-eventstore:2113 \
    -e CLUSTER_MODE=single \
    -e EVENTSTORE_USER=admin \
    -e EVENTSTORE_PASSWORD=changeit \
    marcinbudny/eventstore_exporter

Configuration

The exporter can be configured with commandline arguments, environment variables and a configuration file. For the details on how to format the configuration file, visit namsral/flag repo.

Flag	ENV variable	Default	Meaning
--eventstore-url	EVENTSTORE_URL	http://localhost:2113	Eventstore HTTP endpoint
--eventstore-user	EVENTSTORE_USER	(empty)	Eventstore user (if not specified, basic auth is not used)
--eventstore-password	EVENTSTORE_PASSWORD	(empty)	Eventstore password (if not specified, basic auth is not used)
--cluster-mode	CLUSTER_MODE	cluster	Set to 'single' when monitoring a single node instance, set to 'cluster' when monitoring a cluster. This settings decides whether gossip stats endpoint is queired.
--port	PORT	9448	Port to expose scrape endpoint on
--timeout	TIMEOUT	10s	Timeout when calling EventStore
--verbose	VERBOSE	false	Enable verbose logging
--insecure-skip-verify	INSECURE_SKIP_VERIFY	false	Skip TLS certificatte verification for EventStore HTTP client

Grafana dashboard

Can be found here

Exported metrics

Let me know if there is a metric you would like to be added.

# HELP eventstore_cluster_member_alive If 1, cluster member is alive, as seen from current cluster member
# TYPE eventstore_cluster_member_alive gauge
eventstore_cluster_member_alive{member="172.16.1.11:2113"} 1
# HELP eventstore_cluster_member_is_clone If 1, current cluster member is a clone
# TYPE eventstore_cluster_member_is_clone gauge
eventstore_cluster_member_is_clone 1
# HELP eventstore_cluster_member_is_follower If 1, current cluster member is a follower (only versions >= 20.6)
# TYPE eventstore_cluster_member_is_follower gauge
eventstore_cluster_member_is_follower 0
# HELP eventstore_cluster_member_is_leader If 1, current cluster member is the leader (only versions >= 20.6)
# TYPE eventstore_cluster_member_is_leader gauge
eventstore_cluster_member_is_leader 0
# HELP eventstore_cluster_member_is_master If 1, current cluster member is the master (only versions < 20.6)
# TYPE eventstore_cluster_member_is_master gauge
eventstore_cluster_member_is_master 0
# HELP eventstore_cluster_member_is_readonly_replica If 1, current cluster member is a readonly replica (only versions >= 20.6)
# TYPE eventstore_cluster_member_is_readonly_replica gauge
eventstore_cluster_member_is_readonly_replica 0
# HELP eventstore_cluster_member_is_slave If 1, current cluster member is a slave (only versions < 20.6)
# TYPE eventstore_cluster_member_is_slave gauge
eventstore_cluster_member_is_slave 0
# HELP eventstore_disk_io_read_bytes Total number of disk IO read bytes
# TYPE eventstore_disk_io_read_bytes gauge
eventstore_disk_io_read_bytes 20480
# HELP eventstore_disk_io_read_ops Total number of disk IO read operations
# TYPE eventstore_disk_io_read_ops gauge
eventstore_disk_io_read_ops 2814
# HELP eventstore_disk_io_write_ops Total number of disk IO write operations
# TYPE eventstore_disk_io_write_ops gauge
eventstore_disk_io_write_ops 4421
# HELP eventstore_disk_io_written_bytes Total number of disk IO written bytes
# TYPE eventstore_disk_io_written_bytes gauge
eventstore_disk_io_written_bytes 2.6918912e+08
# HELP eventstore_drive_available_bytes Drive available bytes
# TYPE eventstore_drive_available_bytes gauge
eventstore_drive_available_bytes{drive="/var/lib/eventstore"} 5.6815230976e+10
# HELP eventstore_drive_total_bytes Drive total size in bytes
# TYPE eventstore_drive_total_bytes gauge
eventstore_drive_total_bytes{drive="/var/lib/eventstore"} 6.2725787648e+10
# HELP eventstore_process_cpu Process CPU usage, 0 - number of cores
# TYPE eventstore_process_cpu gauge
eventstore_process_cpu 0.08
# HELP eventstore_process_cpu_scaled Process CPU usage scaled to number of cores, 0 - 1, 1 = full load on all cores (available only on versions < 20.6)
# TYPE eventstore_process_cpu_scaled gauge
eventstore_process_cpu_scaled 0
# HELP eventstore_process_memory_bytes Process memory usage, as reported by EventStore
# TYPE eventstore_process_memory_bytes gauge
eventstore_process_memory_bytes 1.19267328e+08
# HELP eventstore_projection_events_processed_after_restart_total Projection event processed count after restart
# TYPE eventstore_projection_events_processed_after_restart_total counter
eventstore_projection_events_processed_after_restart_total{projection="$by_event_type"} 0
# HELP eventstore_projection_progress Projection progress 0 - 1, where 1 = projection progress at 100%
# TYPE eventstore_projection_progress gauge
eventstore_projection_progress{projection="$by_event_type"} 1
# HELP eventstore_projection_running If 1, projection is in 'Running' state
# TYPE eventstore_projection_running gauge
eventstore_projection_running{projection="$by_event_type"} 1
# HELP eventstore_queue_items_processed_total Total number items processed by queue
# TYPE eventstore_queue_items_processed_total counter
eventstore_queue_items_processed_total{queue="index Committer"} 54
# HELP eventstore_queue_length Queue length
# TYPE eventstore_queue_length gauge
eventstore_queue_length{queue="index Committer"} 0
# HELP eventstore_subscription_connections Number of connections to subscription
# TYPE eventstore_subscription_connections gauge
eventstore_subscription_connections{event_stream_id="test-stream",group_name="group1"} 0
# HELP eventstore_subscription_items_processed_total Total items processed by subscription
# TYPE eventstore_subscription_items_processed_total counter
eventstore_subscription_items_processed_total{event_stream_id="test-stream",group_name="group1"} 24
# HELP eventstore_subscription_last_known_event_number Last known event number in subscription
# TYPE eventstore_subscription_last_known_event_number gauge
eventstore_subscription_last_known_event_number{event_stream_id="test-stream",group_name="group1"} 23
# HELP eventstore_subscription_last_processed_event_number Last event number processed by subscription
# TYPE eventstore_subscription_last_processed_event_number gauge
eventstore_subscription_last_processed_event_number{event_stream_id="test-stream",group_name="group1"} 19
# HELP eventstore_subscription_messages_in_flight Number of messages in flight for subscription
# TYPE eventstore_subscription_messages_in_flight gauge
eventstore_subscription_messages_in_flight{event_stream_id="test-stream",group_name="group1"} 0
# HELP eventstore_tcp_connections Current number of TCP connections
# TYPE eventstore_tcp_connections gauge
eventstore_tcp_connections 1
# HELP eventstore_tcp_received_bytes TCP received bytes
# TYPE eventstore_tcp_received_bytes gauge
eventstore_tcp_received_bytes 17237
# HELP eventstore_tcp_sent_bytes TCP sent bytes
# TYPE eventstore_tcp_sent_bytes gauge
eventstore_tcp_sent_bytes 3423
# HELP eventstore_up Whether the EventStore scrape was successful
# TYPE eventstore_up gauge
eventstore_up 1

Changelog

0.8.1

• FIX: in some cases /stats endpoint scrape will fail due to missing Accept header - see #13

0.8.0

• FEATURE: support for ES 20.6 - see #12
• FEATURE: option to ignore invalid certificates on HTTPS connection - see #11
• FIX: moved to stateless metrics, that should fix the problem (without a workaround) of zombie metrics after projection / member / subscription has been removed

0.7.0

• FIX: for items of variable count (queues, drives, projections, subscriptions, members) the exporter should not return items that have been removed (see #7)

0.6.0

• FEATURE: add HTTP Basic auth to support EventStore 5.0.2+ (see #6)
• FIX: when status code of http call does not indicate success, the exporter will now log a message and it won't report metrics

0.5.0

• FEATURE: new metrics for detecting cluster node status: eventstore_cluster_member_is_slave and eventstore_cluster_member_is_clone

0.4.0

• FEATURE: persistent subscription metrics

0.3.0

• FEATURE: added drive metrics

0.2.0

• FIX: missing eventstore_process_memory_bytes metric
• BREAKING: eventstore_projection_events_processed_after_restart metric renamed to eventstore_projection_events_processed_after_restart_total to comply with Prometheus metric naming rules

0.1.1

• experimenting with dockerhub tags

0.1.0

• Initial version