But they are only useful if you're able to make sense out of them.
Most important for that: using log levels consistently!
· ℹ️ Trace
· 💡 Debug
· ⚪️ Info
· 🟠 Warn
· 🔴️ Error
· ⚡️ Fatal
A small thread about when to use what 🧵↓
(1/7) ℹ️ Trace
Your most verbose logs containing the most fine-grained information.
It gives you detailed insights into what's happening - not only in your code but also in third-party libraries.
Can go as far as documenting every step in a single algorithm.
(2/7) 💡 Debug
Less information than 'trace' level, but still extended to a way that's needed to troubleshoot problems in detail.
Majorly used for pre-production/testing environments and often logs out sensitive information that can't be logged on production.
(3/7) ⚪ ️Info
The base level of logging.
These logs help you to identify the certain states the application or process has gone through, e.g. "User successfully authenticated" or "Payment information has been processed".
There's no granularity rule, but don't overdo it.
(4/7) 🟠 Warn
Something unexpected happened.
It doesn't implicitly mean that a process has failed or the user received an error.
It's a warning about an expected state and further actions that have to be taken, but not immediately.
(5/7) 🔴 Error
Some business functions may be broken or a dependent system is not available, but your core business processes are still intact.
Mostly it requires immediate action and your users are (partly) affected.
(6/7) ⚡ Fatal
Crucial functions of your application are not working and business processes can't be completed by your users, e.g. because your database is not available or your payment provider is not reachable.
Your business can be at risk and this requires immediate action.
(7/7) Key takeaways
Maybe it feels obvious, but it's often times not for your junior colleagues.
Help them by giving them some guidance so they know which information has which severity.
Stumbling through thousands of logs that can't be easily filtered for severity is painful.
• • •
Missing some Tweet in this thread? You can try to
force a refresh
My most received questions:
"How do I start my cloud engineering journey & what's the right path?"
There's not a single or correct path.
There are only recommendations.
A small recap of mine to get yourself going 🧵↓
(1/7) Pick a cloud provider you're interested in:
· Amazon Web Service
· Microsoft Azure
· Google Cloud Platform
It doesn't matter which one you start with.
Even though they are completely different in some ways, a lot of your learnings will be transferable.
(2/7) Register yourself an account
Yes, you need a credit card, but you don't need to be scared of unexpected or exploding costs.
All of the providers are having a generous free tier, allowing you to test and explore their services.
AWS is growing its number of services at a fast pace.
If we're counting namespaces, the statistics over the past decade are mind-boggling:
· 2013: 25
· 2015: 46
· 2017: 78
· 2019: 182
How to keep up with what's new?
A small thread about sources to keep you up-to-date 🧵👇
(1/4) The AWS News Blog
Guarantees to not miss out on new features or services, but also contains interesting statistics and other insights from AWS itself.
Gets updated very regularly, sometimes several times a day.
If you're only focusing on keeping up with the new capabilities AWS provides, that's your major source.
You'll learn about small or big improvements to existing services, introductions of new ones as well as region expansions.
Scared of getting your hands on at AWS because you're in fear of unexpected costs for the cloud?
You don't have to be!
A small thread about steps you can take to sleep better on your journey to becoming an AWS expert 🧵👇
(1/6) AWS Free Tier #1
AWS grants you a lot of room for experiments for different services each month:
- 1m Lambda Requests
- 25 GB of DynamoDB storage
- 100 build minutes on CodeBuild
... and much more!
Even better: if you've recently created your account, you also have additional free limits for the next 12 months, including for example running a EC2 t2 or t3 micro instance without any costs.