Konrad Kowalski (rootsher)Principal Platform & Reliability Architect101001011100111100000111010100011011110000111000

Latency

posts (12)

  1. The Layers of HTTP in the Browser: From fetch() to QUIC9/9

    From fetch() to the First Byte

    The whole lifecycle in one place: browser policies, cache, connection, wire format, headers and body. Only with that model does the Network panel answer…

  2. CPU in Kubernetes from the kernel up12/12

    Anatomy of a CPU problem: from an understated request to p99 latency

    One incident end to end: a 500m request, no limit, no throttling, and p99 climbing from 35 to 180 ms. The whole path of the mechanism.

  3. CPU in Kubernetes from the kernel up11/12

    Noisy neighbor and CPU overcommit: when one Pod breaks another

    Overcommit works until bursts become correlated. An understated request is not only about placement - it is a weaker position under contention.

  4. The Layers of HTTP in the Browser: From fetch() to QUIC8/9

    HTTP/3 and QUIC: Same HTTP, Different Transport

    QUIC removes transport-level head-of-line blocking and sets up a session more cheaply, but it does not fix the dependencies the frontend creates in its own…

  5. Testing Through Layers: From Code to Infrastructure8/9

    Performance Tests and Load Tests: Behavior Under Traffic

    A performance test measures system behavior over time, a load test checks expected traffic, and a stress test looks for the breaking point.

  6. CPU in Kubernetes from the kernel up5/12

    CPU Contention: when a process wants CPU and does not get it

    With no limit at all a process can sit and wait. All it takes is runnable demand exceeding available capacity - and you only see it in latency.

  7. Deployment Strategies: Why They Exist3/6

    Canary Deployment: The New Version Gets Part of the Traffic

    Canary limits blast radius by using a small percentage of real traffic. Without metrics, it is only a slower rollout.

  8. Realtime in the Browser: From Polling to WebTransport7/7

    Polling, Streaming, SSE, WebSocket, or WebTransport?

    The choice follows from the traffic model: freshness, direction, frequency, ordering and reliability. The simplest correct mechanism usually wins.

  9. Realtime in the Browser: From Polling to WebTransport6/7

    WebTransport: Streams and Datagrams for Realtime over HTTP/3

    HTTP/3 and QUIC give a single session many independent streams plus datagrams. Not all data has to arrive reliably and in order.

  10. Realtime in the Browser: From Polling to WebTransport2/7

    Long Polling: When the Server Delays the Response

    The server does not answer until it has something to send. What appears is a chain of long requests, where the timeout and the cursor are part of the protocol.