Loading…

Log tracing for incident response

Follow the tracks to the root cause.

Spoor reads raw logs from any source, draws the infrastructure they describe, and walks your team from the first sign of trouble to the fix. Personal data is masked before anyone, or any model, sees it.

Runs in your browser. Nothing leaves it unless you connect a platform or a model.

From raw lines to recovery

  1. 1

    Bring any logs

    Drop files or whole folders, or pull from Splunk, Datadog, Elasticsearch, Loki, CloudWatch, Google Cloud, Azure or New Relic. Formats are recognised line by line.

  2. 2

    Spoor reads the trail

    It finds the incident window, maps which service calls which, and ranks every warning and error by timing, novelty and reach.

  3. 3

    Your team recovers

    A path from change to recovery, the likely root cause with fixes, where in the world it hurts, and plain-language answers that cite the exact lines.

Built to cut time to recover

Infrastructure map

Built from request ids and messages, with the failing path in red and the root cause outlined. No agent or service registry needed.

Service lanes

Every service against time on one screen. Errors darken, warnings glow, and a service that goes silent shows as a gap.

Root cause, ranked

Five explainable factors, with symptoms such as timeouts discounted so the pattern behind them ranks first. Each cause comes with fixes.

World impact map

Places every failing request from client addresses and country fields, and names the regions with traffic but no failures. No lookup service sees an address.

Status update in one click

Impact, regions, likely cause, trigger, mitigation and next step, written for your incident channel and masked before it reaches the clipboard.

Built for big files

Streams gigabytes in 4 MB pieces across your processor cores. Tested at 1.1 GB and 6.9 million lines in one browser tab.

Safe hands for other people's data

Logs carry customer emails, card numbers, addresses and secrets. Spoor treats every line as sensitive and gives the people handling it guard rails that are on by default.

    Connects to where your logs live

    Files and folders.gz archivesGrafana LokiElasticsearchOpenSearchSplunkDatadogAWS CloudWatchGoogle Cloud LoggingAzure MonitorNew RelicGraylogSumo LogicAny HTTP APIPrivate networks via relay

    See it on a real outage

    The example is a checkout outage across five services. It takes about ten seconds to see what went wrong.

    Where customers are affected

    failing, area = failed requests degraded traffic, no change went quiet

    Infrastructure map

    Service lanes

    no logslogging normallywarningsfew to many errors
    How to read this page
    errorwarnother