Aaaargh deployment failed! ShareFile to the rescue!
Deployment Failed. Log files everywhere! Have no fear, ElectricFlow + ShareFile are here.
WINNER: Best in Show Prize
Laurent and I both work at a company called Electric Cloud. One of our products is called ElectricFlow Deploy and can perform application deployments.
Day to day Laurent and I use Citrix products heavily, so when we saw this hackathon we were excited to learn a few more APIs. At work one of our software distribution mechanisms is via ShareFile and it has served us well. We automated publishing files to ShareFile for customer consumption a while ago. But we wondered, how hard would it be write an integration that gave a generic way to publish files to ShareFile?
There were many use cases where ShareFile can help. In this hackathon entry, we'll focus on the application deployment use case as shown in this flow chart:
When application deployments go wrong you would like to:
- Notify stakeholders
- Track and communicate progress
- Aggregate and share log files to cross functional teams in order to triage deployment failure
- Take follow on steps to fix the failure and prevent it from happening again.
If done manually, proper skill and care is needed to find and share log files. Worse, the person collecting the log files may be under stress due to the business impact of the deployment failure.
We used ZenDesk for items 1 and 2 and ShareFile for item 3. ElectricFlow performed the deployment and served as the orchestration engine.
Our ShareFile integration takes the following parameters:
The application deployment looks like
On error we:
- Gather appropriate log files and zip them
- Upload zip file to ShareFile
- Create a zendesk ticket and comment with the ShareFile URL
In ZenDesk you can see the ticket as follows:
Having an automated way to collect, aggregate, share and track log files makes application deployments slightly less stressful.
ShareFile is great service to hold these log files and tie them to an incident. Not only can these files be used during the deployment failure but also during retrospectives and for future training.
Laurent and I had a ton of fun during the hackathon and we thank the organizers.