Monitoring thousands of server metrics, ready-made for you.
So you won’t miss a thing.

  • When it breaks down
  • Okmeter will notify you
  • You’ll know what’s happening
  • You’ll know what to do
  • You’ll fix it quicker

Detalization that rocks

Okmeter auto-magically collects 100s and even 1000s of detailed metrics about every part of your system so  you won’t miss a thing.

Nginx monitoring

Slow upstream responses, errors, traffic and requests balancing.


Docker monitoring

Containers resource usage, CPU, memory, disk IO, blkio ops, quotas, shares, throttling, for docker, kubernetes and k8s pods

Java / JVM monitoring

JVM heap memory pools usage, off-heap memory, threads and GC counts, GC timings.

php-fpm monitoring

Worker states, resource usage, pools starvation, queue saturation.



Instrument your server application code to get runtime statistics — counters, rates, timers with percentiles and histograms.

MySQL monitoring

Performance metrics, transaction throughput and latencies, slow queries, innodb, Qcache, threads, and locks metrics. Full scans and range joins, tmp tables and buffer pool usage.

MongoDB monitoring

Queries runtime statistics — calls count, in/out traffic per query, mean query times and latency percentiles for each collection and query profile. Mongo errors count.

Postgresql monitoring

Connections, transactions, query timings and resource usage, locks, bgwriter, wal, xlog, indexes, seq scans, replication, autovacuum.


Cassandra monitoring

Request latencies, table ops latencies, cluster nodes errors and unavailabilities, cache stats, commit log, compaction, hinted handoff, read repairs, memtable stats, sstables and disk usage.

Elasticsearch monitoring

Сluster health, indices, replicas and shards, per index stats — size, docs count, segments, ops counts and latencies, fielddata cache and evictions.

RabbitMQ monitoring

Message publishing and delivery, acknolegments, memory usage by queues and messages counts.

Redis monitoring

Performance, memory use, commands call counts and cpu usage. Channels, patterns, connections, rdb bgsave, aof rewrites.

PgBouncer monitoring

Сonnection pool metrics, traffic received and sent, xact counts and times, query and wait times. Client states — active and waiting.


Memcached monitoring

Мemory use, hits, misses, evictions, reclaimed, slabs fill percent, cached items count, memcached commands, used chunks, hit rate, miss rate.

ZooKeeper monitoring

Nodes states, leader elections, outstanding requests, znodes, watches, ephemerals.


Detailed and meaningful DevOps

Why okmeter is perfect for you

Right metrics out of the box

To know what’s happening, you must collect proper parameters. We collected other people’s experience of how things might go south. And we figured out what metrics to collect to get all the info you need for troubleshooting.

Auto-detect of common pitfalls

Okmeter has a large knowledge base of typical problems that regularly occur with commonly-used technologies. Okmeter will automatically run 100s of diagnostic checks for each of your servers and subsystems.

Auto-magical integration

Okmeter will automatically detect every service, process and technology in your cluster and collect all needed metrics. It’ll appear organized in meaningful chart dashboards. All that with no configuration.

Cluster overview with drill-down

Okmeter metrics engine allows combining 1000s of metrics in one chart to get an overview of your whole cluster. And still you can drill-down to any specific subset you want. That allows you to pinpoint perf issues and root causes easily.

Some of our clients

Start free trial

Usefulness that matters

Expert level chart dashboards will guide you through what's going on with your DB, service, server.

Clear view into
PostgreSQL performance

To know what’s happening, you need to collect proper parameters. We collected other people’s experience of how things might go south. And we figured out what metrics to collect to get all the info you need for troubleshooting.


Keep track of every user request with access.log monitoring. Pinpoint slow backend responses and unstable pages.

Detailed resource utilization

Beware of what's happening with your hardware. Whether RAID performance is poor or the CPU is overheated — you'll be notified.

Detailed process monitoring

Easily detect when processes behave badly. Okmeter accounts for CPU, memory, disk io, swap usage of each process.

How it works

Install okmeter agent on your servers

Okmeter agent detects every process running on a server and auto-configures all needed plugins.

It will collect, process and send metrics to okmeter.

Preset expert level dashboards for every piece of your stack. These charts will tell you the full story.

Everything is under control now!

Every minute, okmeter checks a whole bunch of triggers. And you'll get paged if something’s wrong.

Start free trial


Free trial. No credit card required.

per month / per server
  • Only Basic metrics*
  • 1 week data retention
  • email notifications
per month / per server
  • 1 week data retention
  • SMS/email notifications
  • Email support
per month / per server
  • 2 month data retention
  • SMS/email notifications
  • Slack integration
  • Email support
Available on-premises
  • 1 year data retention
  • Slack, Jira and custom integrations
  • SAML single sign-on
  • 24/7 support

* Basic metrics: cpu, ram, disc, network, processes

Start free trial

No credit card required


We have a lot of servers, with virtualization and/or containers, how is it billed?

Okmeter agent is installed on physical as well as virtual servers. It will collect the data on everything happening on and with the server with no regard if it's virtual or not. That's why each server with agent installed will be billed as full unit.
Containers are monitored from the host server, so it won't increase the bill.
If you happen to have more than a hundred of servers — contact us, to discuss the price.

We run on AWS EC2 with Auto Scaling and stuff. Will it work?

In order for okmeter agent to play nicely with Auto Scale, you should install it into the image or setup agent installation in startup/cloud-init scripts.
Okmeter billing works by 95% burstable method, i.e. for each month 5% of the time with the maximum number of hosts is ignored.
We understand that it might be not very suitable in some cases and we're commited to do something better with it. Meanwhile, please, contact us, to discuss possible solutions.

Do you have more questions? Ask us anything.