I recently attended the excellent Manchester Docker Meetup hosted by Disney Streaming Services (formerly known as the lovely folks from Cake Solutions). Firstly, I would like to thank Anirvan Chakraborty, Laura Bria and Eslam ElHusseiny for hosting yet again! Now let’s recap on the event…
A quick shake down of what Logz.io do:
- ELK stack in the cloud.
- Log analysis
- Security built into their service
A.I. applied to the logs – this was pretty intriguing. Logz would find answers to potential issues it spots in the logs, Daniel gave the example of an error message which the AI was able to link to an online forum which contained the solution to the error
Daniel gave his talk on Logging docker with the ELK stack. I found the following things most interesting from his talk.
- Issues with consistency. Apps would be logging similar things in slightly different formats. Times and dates are the obvious ones, but also the log statement format would vary from one app to the next. This makes searching difficult.
- Shipping logs from a container into ELK. Several methods were mentioned, including using Filebeats, using a log driver, using Logspout or a logging collector developed by Logz.io, checkout Daniel’s blog for more details
- ELK is the second most downloaded open source software after Linux.
- Elasticsearch is the second most downloaded docker image after NGINX.
- There was recently a large scale ransomware attack against open source ELK stacks which had failed to implement adequate security controls.
- Efficiency in logging takes up a lot of space. There is a need to be precise and not verbose to save money without losing accuracy / effectiveness of the logs.
This was a really engaging talk from the trenches of moving a monolithic app into containers with everyone’s favourite – microservices architecture – thrown in.
Phil took us through the story of the growth of Interact and the need to move to a DevOps format to speed up delivery and continue a high level of support. They are currently shifting from a hosted service either on the clients hardware or their own, to a Cloud solution provided by AWS.
The stand-out parts of this talk for me were:
- Taking a large well established system and moving it to the Cloud is hard!
- Justifying the time it takes to make this move and the upheaval required needs some good persuasion techniques.
- Separating components out into smaller services makes life easier when working with Docker and AWS.
- Monitoring and alerting is very important to stay on top of a system.
Thanks to both of the speakers for providing very useful insights into the world of Docker.
Let us know what you think in the comments.