ajohnstone

Operations architect at photobox

[ajohnstone @ajohnstone]

@simonplend Lol, famous for it in my current gig.

[ajohnstone @ajohnstone]

@simonplend Why do you need to wrap in node, console.log and eval? Seems unnecessary?

Can’t beat jq, such an amazing tool, formatting and manipulation of the output can be done directly.

jq -n ‘. | “\(.field)”’

Can pipe to tsv etc

[ajohnstone @ajohnstone]

@simonplend Good luck, wish you all the best there…

[ajohnstone @ajohnstone]

@simonplend @benpbarnett been quite out of touch with things regarding web dev, looking to write an app and just wondering what’s trending now? I’ve gone down the route of angular/typescript/graphql. Anything you’d recommend looking into?

[ajohnstone @ajohnstone]

@simonplend I use both extensively
a) yaml can be easier to read at a glance
b) I prefer json for processing/parsing
c) other protocols/formats for more efficient storage and processing.

Jq is awesome for manipulating and managing json

[ajohnstone @ajohnstone]

@simonplend 1. Yaml is a superset of JSON.
2. Supports comments
3. multi line strings
4. Anchors and aliases and overrides
5. Payload sizes are larger in yaml with formatting and spacing typically
6.Readability
7. execution/performance differs with encode/decode
8. Yq is inferior to jq

[ajohnstone @ajohnstone]

@DinisCruz @AWSSupport @QuinnyPig @awscloud @AWSSecurityInfo @ben11kehoe @alexbdebrie @pacohope @abhaybhargav @SeanTMalone @brunogarrancho @AlzetteInfoSec @singhprateik @bjohnso5y @darkp0rt @slipdexic a) loggers
```
logging.getLogger('botocore').setLevel(logging.DEBUG)
logging.getLogger('urllib3').setLevel(logging.DEBUG)
```
b) monkey patch boto3/urllib3

This exposes sensitive data, so add a filter on the loggers.

[ajohnstone @ajohnstone]

@DinisCruz 1. Aws lambda has its own scheduler.
2, docker socket would not be exposed.
3. Virtual kubelet could be ran and a rootless docker ( https://t.co/FU04mZfsSO )

So maybe possible.

[ajohnstone @ajohnstone]

@DinisCruz Statectr could be messed up. See code here, repeatable after a daemon restart?

https://t.co/tAlCivw7O9

https://t.co/cKKdJBEaqC

[ajohnstone @ajohnstone]

@simonplend Curious about the format your using, mistakes, knowledge and skills and how that helps you arrive at a product?

Not seen that before, would of thought some kinds of mind map?

[ajohnstone @ajohnstone]

@hillelogram I once wrote a backend based on google sheets, so I didn’t have to write a backend. Just make sure you use protected fields/columns and use validation on input :)

[ajohnstone @ajohnstone]

@DinisCruz It’s late, start fresh in the morning 🥴

[ajohnstone @ajohnstone]

@DinisCruz Your permissions will be incorrect for your iam role you have assumed. It won’t be an issue with AWS.

[ajohnstone @ajohnstone]

@simonplend I also use warning to add important information that can easily been seen above informational messages to add emphasis on the message.
It does not necessarily mean an error occurred.

[ajohnstone @ajohnstone]

@ejc3 @DinisCruz @Rancher_Labs Rancher used to have issues at scale, and looks like it’s improved now. They quote the following

> GA – support for 2000 clusters, 100,000 nodes
and quotes 5k nodes

https://t.co/DT5Pu3YlRB

[ajohnstone @ajohnstone]

@brianleroux @DinisCruz Would probably create a sqs queue, and trigger aws lambda off the back of it. Entirely depends how your processing the files.

[ajohnstone @ajohnstone]

@DinisCruz Lol, allergic and java in the same sentence ;)

[ajohnstone @ajohnstone]

@DinisCruz If it was bubbled up it would be raised and then handled :)

[ajohnstone @ajohnstone]

@DinisCruz This demonstrates bad practice.
1. Do not wrap and hide exceptions
2. Bubble up and handle the exceptions correctly in the context they are used.
3. Unfortunately, above demonstrates to engineers how to suppress exceptions.
4. I cannot count the amount of times I’ve rejected PRs