Logmatic.io as a daily tool to improve IT operations
We recently met with George Plancke, CTO and co-founder of SimpliField, one of the early customers of Logmatic.io. He looked quite a happy man and so we asked him a few questions we thought we’d share with you.
Hi Georges – can you tell us a bit more about SimpliField?
We improve field operations and performance in the retail world. Our customers use SimpliField to empower their employees to create awesome field reporting in real-time & lead analysis on it. The results are very impactful: one of our customers, a large telco operator has gained 6 weeks in its decision-making process regarding its fast-moving retail store business.
So, how did you meet Logmatic.io?
When the Logmatic.io team contacted me to demo their product, I was already in the process of looking for a specialized log analytics solution for Simplifield. I wanted to understand what my customers were doing on our system and monitor the health of my infrastructure. Logmatic.io sounded like having a powerful but simple approach: it would allow SimpliField to have one monitoring tool for nearly everything. Moreover, there are no complicated frameworks to integrate nor any external services pollution in our product. We just send our logs and with the dashboards we built inside Logmatic.io’s platform, we get everything we need. Their vision really appealed to me.
Before working with Logmatic.io, what was monitoring & troubleshooting like at SimpliField?
Before working with Logmatic.io, we were troubleshooting almost exclusively with the famous tail linux command on each machine of our infrastructure. We had of course New Relic, but this service has proven itself to only be dashboard sugar with no real use for real and serious debugging. With the Logmatic.io log analyzer, we have those graphics perfectly tailored to our needs and the source logs from all our machines making troubleshooting a breeze (well, far less painful anyway).
How much time did you spend setting it up?
Setting up Logmatic.io was really easy, It was just a matter of a few lines in Syslog-ng. We thought for a little while to develop a custom implementation for Winston. But then we decided it was simpler to keep all our logs redirected to Syslog. It took literally 15 minutes to get up and running. Well, 1h if you count the time spent tweaking our log format a little bit to have richer data analysis.
And what are the results?
As many devops/sysadmins, we thought command line was (and often is) all powerful, and there is nothing you can’t do with it. But it has not been designed for the purpose of troubleshooting: it’s clearly not the job best done with CLI. Once you get used to a well-thought GUI (with the extra benefit of being in the cloud), there is no way back.
Before using Logmatic.io, it was always the same ritual, opening multiple secured shells, tailing -f a couple of log files trying to find out little by little what could have been wrong.
With its powerful search, the fact that all our clusters’ logs are sent to it, we can more easily find correlations betweens processes, spot the culprit quicker (fail in load balancer, in the app, the database ?).
For an order of magnitude, I think it takes 3-4x less time than before, but more importantly, the alerting system allowed us to often detect problems before it becomes really hazardous for our business.