Prometheus Pull vs Push: A Tradeoff, Not a Moral Position
Scrapes, service discovery, Pushgateway, and OpenTelemetry collectors: when pull is the operational win, and when short-lived jobs need push.
Category
33 articles · 119 min of reading
Scrapes, service discovery, Pushgateway, and OpenTelemetry collectors: when pull is the operational win, and when short-lived jobs need push.
bpftrace, BCC, and production agents: kprobes versus tracepoints, map memory, and why CO-RE beat copying kernel headers into the cluster.
How Alibaba's Aone platform centralized build, change, and security gates so thousands of applications could ship without each business unit owning a unique CI snowflake.
Which of Heroku's 2011 factors still prevent outages, which ones aged (local disk, log sockets), and how to apply them without cargo-culting config vars.
Horizontal versus vertical autoscaling in Kubernetes, custom metrics, and the cooldown and request-sizing mistakes that cause flapping.
Lead time, deploy frequency, change fail rate, and time to restore — how to collect them honestly, and how teams game the dashboard.
What SLSA levels actually require, how provenance attestations fit next to SBOMs, and a realistic path from 'we tag git' to a verifiable builder.
GitHub-hosted VMs versus autoscaled ephemeral self-hosted: dirty workspaces, credential lifetime, and when Buildkite's similar model is the same idea.
Flakes, nix-shell in GitHub Actions, binary caches, and the difference between bit-reproducible artifacts and 'it built on Tuesday'.
sync-wave annotations, hooks, and waves that wait: how to roll CRDs before CRs, and the deadlock of a wave that never becomes healthy.
traceparent, baggage, SDK propagators, and the one missing middleware that turns a mesh into a pile of orphan spans.
How lockfiles, dependency audits, and software bills of materials work together to manage the risk of running someone else's code in production.
Monorepos and polyrepos both work at scale — the real decision is which trade-offs your organization is already equipped to absorb.
Release flags, experiment flags, and ops flags need different lifecycles — a flag system without a removal process always accumulates debt.
How to set service level objectives and error budgets that drive real engineering decisions, instead of becoming a number nobody looks at after launch.
A measured look at which Spring Boot startup optimizations move the needle in practice, from lazy initialization to Class Data Sharing and beyond.
Logs, metrics, and traces are usually taught as separate pillars, but the real value comes from how they connect during an actual incident investigation.
The testing pyramid was always a proxy for a cost trade-off — here's the trade-off itself, and when the pyramid's shape doesn't apply.
A concrete comparison of blue-green and canary deployment strategies, including the infrastructure they each require and where one clearly beats the other.
A comparison of secrets management patterns, from environment variables to dedicated vaults, and how to pick the right one without overengineering a simple app.
Why Terraform state is the most fragile part of most infrastructure setups, and the locking, backend, and workspace patterns that keep it from becoming a liability.
How GitOps replaces imperative deploy scripts with a reconciled, auditable source of truth, and the practical patterns for running it well with Argo CD or Flux.
How trunk-based development and merge queues reduce integration pain compared to long-lived feature branches, and what they require to work in practice.
Practical caching techniques for CI pipelines, from dependency caches to Docker layer caches, and how to avoid the stale-cache bugs that undo the savings.