What tools do you use for monitoring?

I'm using AWS Cloudwatch it's been a good enough solution but we want to take it to the next level by using something like ELK. We also want to connect the AWS alerts to a Slack channel (at the moment we use email).

Server: NewRelic
Logging: Timber
Marketing/Usage: Segment (With GA and a bunch of other tools attached)
App bug reporting: Sentry
Backend bug reporting: Appsignal

Honestly, we don't monitor or log 10% of what we should. Just haven't had the time yet to implement better monitoring. #StartupLife

I have used the following:

• CloudWatch Logs/Metrics/Alerts: it's ok but dashboards are not super pretty, querying logs is not super easy and not fast.
• ELK + elastalert: pretty cool dashboards, easy to query data. Cons: easy to overload you ElasticSearch instance with data, type conflicts, logstash problems
• CloudWatch Logs + Lambda + ElasticSearch + Kibana: you don't have logstash but the rest remains
• New Relic, quite neat, powerful query language, nice dashboards, lots of plugins. Cons: probably expensive

I'll just focus this response on client-side and Node error handling. For errors we don't handle, we've been relying on Sentry. We've been pretty happy with it. I've also heard good things about Track JS.

When I was still doing .NET, we relied on the Enterprise Library Application Blocks for logging and exception handling.

In my company and my personal projects, I use:

• Sentry: online version or selfhosted, for error handling. It does a great job capturing exceptions, showing all the exception trace and not duplicating it if it happens more than one time.
• Jenkins: monitoring tests and staging server deployment output.
• Zabbix: monitoring almost all other things: servers availability via ping, server processes uptime and performance with templates, backups output, web sites response code and time, periodic background processes being executed, etc.

Riemann (riemann.io/) is pretty exciting. It's awesome because the config is written in Clojure. It's terrible because the config is written in Clojure.

It's very fun to use, and it supports unit testing your config which is very nice. It's not something I'm using right now as I prefer SaaS to running my own services, if at all possible, but I have used it successfully in the past couple of years.

We are using the following tools:

• cloudwatch for the aws setup
• ELKstack in combination with logstash for log aggregation of all tools
• CheckMK (mathias-kettner.com/check_mk.html) to monitor more deeper details for every host + we started integrating app specific checks as well

NewRelic is nice to have but extremely expensive when you are running > 60 servers in production, using it for only a few hosts does not help you in my eyes.

At my company, we use New Relic which monitors almost everything. And if you want something specific to Rails applications then even skylight.io is a good option!

Icinga2 for service monitoring with email and sms alert (for important services).

InfluxDB/Grafana for performance monitoring, data comes mostly from icinga but some from legacy things from collectd.

We're using ELK with variety of Filebeats connected. It's pretty convenient, cause you can easily plug in and out new elements. And dockerized version is an opportunity to upgrade versions without excessive overhead