Monitor the services running next to your workloads.
WatchLog finds databases, caches, queues and proxies in what its agents already see: in a Kubernetes cluster, in Docker, or installed on the host of a Linux server. Connect one with the command it writes for you, or none, and get a full page, default alerts and plain-language reasons.
RedisRedis 6.2 and later, and ValkeyMemory against maxmemory, evictions, hit rate, the slow log, clients and replication. Flags KEYS in the slow log and a missing maxmemory.MemoryCommandsClientsPersistence and replicationSetup: none without a passwordLearn more about Redis →
PostgreSQLPostgreSQL 12 and laterConnections against the limit, the slowest queries as normalised text, who blocks whom, replication lag, estimated bloat and vacuum, transaction ID age.ConnectionsQueriesLocks and waitsReplicationTables and storageVacuumSetup: one commandLearn more about PostgreSQL →
MySQL and MariaDBMySQL 8.0 and later, MariaDB 10.6 and later, and Percona ServerConnections against max_connections, the slowest statements, who blocks whom with the session holding the lock, InnoDB, replication and the largest tables.ConnectionsQueriesLocksInnoDB and storageReplication
Setup: one commandLearn more about MySQL and MariaDB →
MongoDBMongoDB 6.0 and later, standalone or a replica setOperations and latency, replica set members and lag, slow operations with the client that sends them, the WiredTiger cache and the largest collections.OperationsCache and storageSetup: one commandLearn more about MongoDB →
NginxNginx, OpenResty and ingress-nginxRequests, 5xx and latency per upstream from eBPF, with nothing to set up; ingress-nginx's own metrics when they're on, and connections from stub_status.ErrorsLatencyConnectionsSetup: noneLearn more about Nginx →
TraefikTraefik 2.10 and later, and 3.x; k3s's bundled Traefik tooRequests, 5xx and latency per entrypoint and service, TLS included, the backends with no server up, and certificates, from Traefik's own metrics. Its numbers join the service map.ServicesEntrypoints and TLS
Setup: none (metrics on)Learn more about Traefik →
RabbitMQRabbitMQ 3.10 and laterQueue depth, consumers, publish and deliver rates, memory and disk alarms. When a queue stops draining, it names the consumer that was probably lost.QueuesMessage ratesResourcesSetup: none (plugin on)Learn more about RabbitMQ →
KafkaKafka 3.0 and later, KRaft or ZooKeeperBrokers up, offline and under-replicated partitions, consumer-group lag per topic, and the group that lost its members, read through the admin API: no JMX agent.TopicsConsumer groupsBrokersBroker internals
Setup: none on PLAINTEXTLearn more about Kafka →
PHP-FPMPHP-FPM with the status page onBusy and idle workers per pool, requests waiting, pm.max_children reached, requests a second and request time, from each pool’s status page.OverviewRequests
Setup: two lines in the pool’s fileRead the docs for PHP-FPM →
Coming next
ElasticsearchPlanned- MinIOPlanned
- Request an integration →
How integrations work
- 01
Found automatically
The agents already see each container's image and ports, and each service installed on a server's host. WatchLog lists what it recognises on the Applications page; press Monitor on the ones you want.
- 02
Connected with the command it writes
When a read-only user is needed, WatchLog writes the exact command with your names and a generated password. Or keep the password yourself, in a Kubernetes Secret or a file on the Docker host.
- 03
A page, alerts and reasons
Every service gets a full page, default alerts and sentences that explain problems, with the workloads and containers that use it.
Read-only by construction: a fixed list of statements and commands for each kind of service, and monitoring roles only. On Linux servers the install runs the integrations collector next to the agent; with docker run agents, the service page gives its command. On Kubernetes the chart already runs it. What the collector can do.