Aidbox Metrics Server
Define
BOX_METRICS_PORT
environment variable with monitoring server port number.Aidbox starts monitoring server on startup automatically.
To check the monitoring server works make
GET <AIDBOX_BASE_URL>:<BOX_METRICS_PORT>
request. The output should be a string "aidbox metrics".There are three types of metrics Aidbox collects and exposes. All endpoints are available on a separate port, e.g.
GET <AIDBOX_BASE_URL>:<BOX_METRICS_PORT>/metrics
.Endpoint | Update frequency |
---|---|
GET /metrics | continuous |
GET /metrics/minutes | every minute |
GET /metrics/hours | every hour |
The
/metrics/hours
response can take some time since it collects a lot of information from the database. Make sure your metrics scraper timeout is sufficient.global:
# omitted global configuration values
external_labels:
monitor: 'aidbox'
scrape_configs:
# omitted default scrappers configuration
- job_name: aidbox
honor_labels: true
scrape_interval: 10s
metrics_path: /metrics
static_configs:
- targets: [ 'aidbox.example.com:9999' ] # should be <AIDBOX_BASE_URL>:<BOX_METRICS_PORT
- job_name: aidbox-minutes
honor_labels: true
scrape_interval: 1m
metrics_path: /metrics/minutes
static_configs:
- targets: [ 'aidbox.example.com:9999' ] # should be <AIDBOX_BASE_URL>:<BOX_METRICS_PORT
- job_name: aidbox-hours
honor_labels: true
scrape_interval: 10m
scrape_timeout: 30s # increased timeout
metrics_path: /metrics/hours
static_configs:
- targets: [ 'aidbox.example.com:9999' ] # should be <AIDBOX_BASE_URL>:<BOX_METRICS_PORT
Metric | Update frequency | Description |
---|---|---|
aidbox_http_request_duration_seconds_bucket | continuous | request duration as cumulative counters for buckets |
aidbox_http_request_duration_seconds_count | continuous | request duration events count |
aidbox_http_request_duration_seconds_sum | continuous | sum of request duration events value |
aidbox_http_request_wait_seconds_bucket | continuous | queue waiting time as cumulative counters for buckets |
aidbox_http_request_wait_seconds_count | continuous | queue waiting time events count |
aidbox_http_request_wait_seconds_sum | continuous | sum of queue waiting time events value |
Metric | Update frequency | Description |
---|---|---|
pg_requests_total | continuous | number of executed selects requests |
pg_inserts_total | continuous | number of executed insert statements |
pg_updates_total | continuous | number of executed update statements |
pg_deletes_total | continuous | number of executed delete statements |
pg_blks_hit | continuous | number of shared block cache hits |
pg_blks_read | continuous | number of shared blocks read |
pg_tup_fetched | continuous | fetched tuples number |
pg_tup_returned | continuous | returned tuples number |
pg_errors_total | continuous | number of errors |
pg_activity_count | continuous | number of PG workers |
pg_idx_scan | every minute | number of index scans |
pg_seq_scan | every minute | number of sequential scans |
pg_stat_statements_total_calls | every minute | number of times executed |
pg_stat_statements_stddev_execution_time | every minute | statement execution time |
pg_stat_statements_mean_execution_time | every minute | mean statement execution time |
pg_table_size | every hour | table size |
pg_database_size | every hour | database size |
pg_activity_max | every hour | maximum number of connections |
Metric | Update frequency | Description |
---|---|---|
hikari_active_count | continuous | number of active connections |
hikari_idle_count | continuous | number of idle connections |
hikari_acquire_created_seconds_bucket | continuous | time taken to create an actual physical connection |
hikari_acquire_created_seconds_count | continuous | number of created physical connections |
hikari_acquire_created_seconds_sum | continuous | total amount of time to create all physical connections |
hikari_acquired_total | continuous | number of obtained connections |
hikari_acquire_wait_seconds_bucket | continuous | time taken to obtain a connection |
hikari_acquire_wait_seconds_sum | continuous | total amount of time to obtain all connections |
hikari_acquire_used_seconds_bucket | continuous | time consumed by a connection |
hikari_max_size | every hour | maximum number of connections |
Metric | Update frequency | Description |
---|---|---|
jvm_gc_time | continuous | garbage collector execution time |
jvm_gc_count | continuous | garbage collector count of launch |
jvm_heap_memory | continuous | heap memory usage |
jvm_non_heap_memory | continuous | non-heap memory usage |
jvm_thread_count | continuous | number of live threads including both daemon and non-daemon thread |
jvm_thread_peak_count | continuous | peak live thread count |
jvm_thread_daemon_count | continuous | number of daemon thread |
jvm_available_processors_size | every hour | number of processors available to the JVM |
jvm_max_memory_size | every hour | maximum amount of memory that JVM will attempt to use |
jvm_total_memory_size | every hour | total amount of memory in JVM |
If you have a different pg exporter you can disable Aidbox PostgreSQL metrics for avoiding metrics duplication.
In this case, you should set
BOX_METRICS_POSTGRES_ON
to false
valueLast modified 5mo ago