Engineering

Increasing efficiency in Technical Support

Here at Hostinger we value our customers time and are always trying to be as efficient as possible.

In the past few weeks, I noticed a pattern of specific inquiries from our Customer Success department:

“Client is getting a HTTP 503 response, maybe you can provide more information?”

Hostinger is using CloudLinux OS for its vast web hosting infrastructure, which provides user isolation and resource management per account basis. Sometimes clients reach the set thresholds on the account (for example, if the account is under a DDoS), and then some subsequent visitors of that account may temporarily receive a HTTP 503 response.

The plain HTTP 503 response may not be very informative to the end user (it can be caused by some other reasons as well, for example, from custom backend script), so I thought, how can we reduce the confusion and provide more clarity? We do have a lot of data from CloudLinux LVE module, it’s just a matter of means to show it to the Customer Success team, and communicating it to the customer.

I found this little plugin for our Slack bot which nicely integrates with our Grafana metrics platform. Installation is pretty straight forward, you just need to follow the instructions from their Github page. It took a few extra steps on our side, since I deployed the plugin via Chef and setup an Amazon S3 bucket for storing the screenshots from Grafana.

Our CS team can access all the metrics just by writing one command to the Slack bot:

To sum up, –

  • Our client’s are getting the information faster.
  • Technical Support team gets a nice graph that they can provide to the end user, which is a lot more intuitive and easier to understand.