 Good morning, my name is Harry Kimple and I work for New Relic And I have been always been a software developer by heart But my personal opinion is that I think we as an industry need to do a much better job at you know You know correlating more data adding more context to the services that we deeply care about and Why do I have this opinion? Well, I think that because observability is all about context right without replying to observability You know we are just maybe monitoring our servers and infrastructure, right? So maybe monitoring the CPU and the memory of our servers and hosts or maybe Monitoring the response time of our applications or microservices But what is the context that I'm referring to right? So well before I dive into a practical example, let me share a true story of my early career So we implemented this application This web application and we allowed, you know, we had a bug in their system Right which allowed the all of the users to access any data of our entire web application Even the paid content that would only be available for subscribers and pay only users, right? So the long story short is that you know the CEO at some point Got so angry that he actually you know came to my office he would literally stay behind my chair and waited for me to To fix that buck and deploy a new fixed release into production So he would literally stay behind me until this is this is fixed. So not a really good experience for me the context some valuable context that I Would maybe have is maybe like a GPS tracker attached to my CEO So that so that I would see him come approaching me right and go to my office So that would maybe be one of the examples of a good context to have a More practical Example is the following Imagine a customer just received his new car and he's trying to enroll his new car in the manufacturer's portal And for some reason that isn't working as expected for the customer So this failure might trigger a chain of events You maybe have similar examples in your cases So the customer will maybe try again, and he may fail again the customer probably At some point calls the hotline or the customer support The support forwards the ticket to the next level of support Who will then at some point get in touch with the developers and ask them to fix the issue? The developers will maybe you know look at the issue try to fix the problem and try to you know solve that issue and That is something where you know the developers get distracted from the day-to-day work right there working Maybe on the next new hot feature that the is so urgently demanded by the business and product owners That they they get distracted so they are not able to continue that that work at some point the Developer maybe or the the customer maybe you know get so angry that he shares his anger on social media And so that marketing needs to get involved and help you know get you know alleviate that Complained by the customer and try to get you know the brand experience Protected so all of this you know It's relevant information For me to have in my day-to-day job right all of this is context that I'm maybe interested in in applying to my services And maybe help me as a software developer, but also help maybe others in the environment to you know dependent upon their role right Another example is probably a little bit more tangible The most common failures are probably service or physical service disruptions So we cannot avoid these even with the best high availability Architecture in mind right we cannot avoid these service disruptions So but what about a service response with unavailability or maybe you know the client did something that it shouldn't have done So personally I would also consider a response duration of larger than the 99s percentile of my entire Responses as a fail as well right and this is a little bit you know Reflected in the so-called apt X score which is an industry score from an observability perspective But maybe you know this issue you know the site can be reached you know There's an error that is happening you know a lot of failure But you know when we agree on the fact that some 4xx errors are normal We might also agree that some you know 4xx responses shouldn't be there at all So I trust that the 404 issues are kind of like a low hanging fruit right to solve But the cost of them and their context related to infrastructure budget is Quite significant and this may be important information to gather and take into account when we look at all the infrastructure So taking also into account the cost perspective of our services as part of the context that we collect and This may lead to you know like I said significant infrastructure cost So these examples you know they should hopefully inspire you to think about ways on how you could improve Your visibility into your specific use cases The real superpower of observability in my opinion is the context that hopefully brings to life And that is why I do have the strong opinion that context is really king And that is the real superpower of observability at the end of the day Thank you very much