LIST
platform team should provide the capability that others use. monitoring team doesn't do the monitoring it provides the ability for others to do their own monitoring shift from ownership of the domain to offering a service to customers (src: Book_ release it! - Michael Nygard)
push push over network 🙂 containers no long-lived identity no local storage pull central machine reaches out to all known hosts services write to local files (src: Book_ release it! - Michael Nygard)
small set of instances that get the new build first. For a period of time, instances running the new build coexist with instances running the old build If the canary instances behave oddly, or their metrics go south, then the build is not rolled out to the remaining population todo: ref handling versions - peaceful side-by side running of different versions (src: Book_ release it! - Michael Nygard)
Don’t change runtime instances Destroy and recreate Always start from a stable, known base image (src: Book_ release it! - Michael Nygard) (src: Evolutionary Architecture - 9. Guidelines for Evolutionary Architecture)
Pattern: put all dependencies in a private repository Code (eg. Maven) Pipeline (eg. Jenkins plugins) (src: Book_ release it! - Michael Nygard)
Pattern: split deployment view and runtime view Deployment view Transform sources into binaries Transform binaries into deplotments Runtime view Processes running on machines (src: Book_ release it! - Michael Nygard)
Stability Pattern_ Shed load turn away load we can't complete in time Where? Load balancers How? Health check + SLA a good health check on the first tier of services can inform the load balancer when response times are too high (in other words, higher than the SLA) Model_ health check Own response time services can measure their own response time to help with this. They can also check their own operational state to see if requests will be answered in a timely fashion....
(Book_ This Is Lean - Niklas Modig) TL;DR Lean means valuing flow efficiency over resource efficiency Choosing for lean operations is a strategic choice with pros and cons. Maybe also Lean means valuing continuous small improvements over one big improvement. That continuous improvement is aimed towards the (impossible) ideal of 100% flow efficiency AND 100% resource efficiency. But favors flow efficiency, so it will focus first and focus more on flow efficiency....
Migration Challenges (1) migration costs + (2) shifting to opex (3) cloud readiness (4) entrenched IT org stakeholders infra cost: perceived vs actual const saving vs pricing engagement process extra: TSO Logic: 3rd party cloud assessment migration portfolio assessment tool (MPA) Migration Challenges (1) migration costs + (2) shifting to opex (3) cloud readiness (4) entrenched IT org (1) migration costs + (2) opex model sunk cost questions:...
x-axis: horizontal duplication scale by cloning eg. load-balander + auto-scaling y-axis: functional decomposition scale by splitting different things eg. monolith -> microservices (compare with Model_ 5 service disintegrators #3 scalability and throughput) z-axis: data partitioning scale by splitting similar things “sharding” eg. split data per product-id (shard 1 = product-id 0 - 99, shard 2 = product-id 100 - 199, …) eg. split data per time (aka “archiving older data”)...