Menu

Build StaaS with microservices: The pipeline metrics (Part 3)

gaming on computer
Table Of Contents

Share this page

Tony Johnson
Guest Author: Tony Johnson, SRE Automation Lead, IBM

In my last blog, I talked about designing and executing the storage pipeline we created at Red Hat using Ansible modules for NetApp® technology. Next, I’d like to share the results we saw from undertaking this journey.

Before I talk about the KPIs and metrics we measured, I’d like to review a couple of cultural changes I observed, which are just as important.

The first thing I noticed was that the team was less defeated, more engaged, and happier. They contributed ideas and solutions in group and organizational meetings. They were also more upbeat with automation and started to develop an automation mindset. I noticed that as the team got requests, they would discuss ways to automate tasks so that they wouldn’t have to do them manually.

The team started coming up with customer-focused solutions without management direction. An example of this was a customer self-service dashboard, where application owners could do an initial storage triage without the storage team. The dashboard would display the IOPS and latency SLAs, and also the current applications’ storage performance. This significantly reduced complaints around application storage performance. It would give advice like “If you observe the A scenario, contact network and compute. If you observe the B scenario, contact us. We’re here to help.”

Let’s dig into the metrics

As displayed in the following graphic, we saw a tremendous increase in speed and efficiency, and a simultaneous increase in quality metrics. Our provisioning cycle time went from 5 days to 5 minutes. Provisioning errors reduced from 1 error per 75 requests to 1 error per 300 requests. Because the automation error was typically a Management, Instrumentation, and Discovery (MID) Server being down or some other infrastructure not being available, the mean time to recover from errors went down from 4 hours to 20 minutes.

red hat pipeline metrics

Automation takes effort. It can require a culture shift, training, bringing in help, missteps, and organizational pushback to make it work. But make no mistake, infrastructure automation is worth the effort. Start by making a plan. See where you want to go. Take small steps on the road. Know how to measure success and failures. Be ready to adjust, and go for it.

Learn more

Learn more about automation with NetApp and Ansible.

Guest Author: Tony Johnson, SRE Automation Lead, IBM

Tony Johnson is the SRE Automation Lead for IBM. He is responsible for the automation of hybrid cloud services and platforms for the IBM CIO team. Previously he was Storage Manager at Red Hat, from 2014 to 2020, leading a group of 6 engineers responsible for IT storage.

View all Posts by Guest Author: Tony Johnson, SRE Automation Lead, IBM

Next Steps