StackStorm Connects DripStat And PagerDuty

June 11, 2015

Guest post by Prashant Deva, Chief Dripper and CTO of Chronon Systems, parent company of DripStat. Earlier this year, Prashant demonstrated automated responses to OutofMemory exceptions and high GC pause times with StackStorm.  

A common request among users of the DripStat APM is to get their alerts inside PagerDuty. Today we will show how using StackStorm, one can connect DripStat to PagerDuty.  So why bother using StackStorm? Well, StackStorm gives you thousands of integrations plus the ability to tie them together with a rules engine, workflow, audit, a GUI and CLI, API, and more.

StackStorm is like an “If This, Then That” but for your operating environment.

Once DripStat informs you that something has gone wrong with the system, you can use StackStorm to automate more and more of your responses to such events and conditions. Thus making a self healing data center.

Install StackStorm

Installing StackStorm is easy.  Just enter the following commands on your console and in a few minutes StackStorm will be fully installed.

curl –q –k –O
chmod +x
sudo ./

st2 auth testu –p testp

The above step will output an auth token. Replace its value in the command below:

export ST2_AUTH_TOKEN=0e51e4d02bd24c2b9dac45e313e9f748

Install The DripStat And PagerDuty Packs

StackStorm has the concept of ‘packs’, which are units of functionality specific to a service. Let’s install the packs from DripStat and PagerDuty since we will be connecting them together.

st2 pack install dripstat,pagerduty

cd /opt/stackstorm

Edit DripStat Pack config:

vi configs/dripstat.yaml

Enter API key from your DripStat Account

Edit PagerDuty config:

vi configs/pagerduty.yaml

Enter the API Key and Service Key in this file.

Create Pack To Connect DripStat To PagerDuty

Now lets create a pack that will hold our script that takes DripStat alerts and forwards them to PagerDuty.

cd packs
mkdir monitoring
cd monitoring
mkdir actions rules sensors

Create file:

vi rules/dripstat_alerts_to_pagerduty.yaml

Enter the following contents:

name: dripstat_alerts_to_pagerduty
description: “Take all incoming alerts from DripStat and automatically notify on-call person via PagerDuty”
enabled: true
type: dripstat.alert
criteria: {}
ref: pagerduty.launch_incident
details: “Alert ({{trigger.app_name}}: {{trigger.alert_type}} on {{trigger.jvm_host}} {{trigger.started_at_iso8601}}”
description: “Alert triggered from DripStat on Application – {{trigger.app_name}} started at {{trigger.started_at_iso8601}}. Affected host: {{trigger.jvm_host}}. Alert message: {{trigger.alert_type}}”

Now just do:

st2ctl reload --register-all

And that’s it. You are good to go!

And More…

Within just a few minutes you now have DripStat alerts going into PagerDuty. You can use the knowledge in this tutorial to connect DripStat with all the other services that StackStorm supports too. For example, you can have it:

  • Create an issue in Jira upon alert trigger
  • Attach the exact github commit of the deploy that caused the alert
  • Alert folks on SMS using Twilio
  • Move your application automatically to a bigger box
  • As explained earlier this year, auto remediate OutOfMemory exceptions and highGC pause times
  • Use ChatOps – StackStorm recently released some capabilities that should make ChatOps easier to manage and extend.

And much, much more. The more you automate, the more robust your infrastructure and your application.  With StackStorm it is simple to implement high degrees of automation without the sense of a giant basket of technical debt hanging over you.

Take a look.  Feedback welcome. Check out DripStat, if you haven’t already.  I’m at @pdeva  – and we can also be reached at @dripstat.

If you have questions about StackStorm there is an active community on Slack at #community, over on IRC on Freenode at #StackStorm.  Or email them on moc.mrotskcatsnull@troppus  or ping them on @stack_storm (note the underscore).