Zipkin: Distributed Systems Tracing at Twitter

Ideas you will meet

Distributed Systems

Consensus, time, failure and the fundamentals of scale-out.

Observability

Metrics, logs, traces, SLOs and alerting.

What we already know about this

Our own lessons on these ideas. Read them here before you start, so it makes sense.

• System Design Foundations · 9 minObservability and running a systemMetrics, logs and traces, how to define 'healthy' with SLOs, and how to ship changes without breaking things.Read hereClose

  You cannot fix what you cannot see

  A system you cannot observe is a system you cannot trust. Observability is the ability to understand what is happening inside from the outside. Three kinds of signal work together.

  Signal	What it is	Best for
  Metrics	Numbers over time (requests per second, error rate, CPU)	Dashboards, alerts, trends
  Logs	Timestamped records of events	Details of what happened
  Traces	The path of one request through many services	Finding where time is spent

  100%

  Figure SDA-FND-11-1 · From alert to root cause · © forcsguy.com

  What to measure

  A simple checklist, the four golden signals:

  1. Latency: how long requests take (use percentiles).
  2. Traffic: how much demand there is.
  3. Errors: how many requests fail.
  4. Saturation: how full the busiest resource is (CPU, memory, connections, disk, queue depth).

  Averages hide pain. If the average is 100 ms but p99 is 4 seconds, one user in a hundred is waiting four seconds, and with many requests per page that is most of your users sometimes.

  Defining healthy: SLIs, SLOs, error budgets

  • An SLI (indicator) is a measurement, like "fraction of requests answered successfully in under 300 ms".
  • An SLO (objective) is the target, like "99.9% over 30 days".
  • The error budget is what is left over: 0.1% of requests may fail. While there is budget, ship features; when it runs out, focus on reliability.

  Alert on symptoms users feel (error rate, latency) more than on causes (one machine's CPU). Page a human only for things that need action now.

  Ship changes safely

  Most outages come from changes. Reduce the risk:

  • Small, frequent releases are easier to understand and undo than big ones.
  • Canary release: send a few percent of traffic to the new version and watch the metrics before going wider.
  • Feature flags: deploy code turned off, then enable it gradually, and switch it off instantly if needed.
  • Blue-green: keep the old version running so you can switch back in seconds.
  • Backwards-compatible changes: database and API changes should work with both old and new code during rollout.
  • Fast rollback must be a rehearsed, boring button.

  When things go wrong

  Have an incident process: someone leads, someone communicates, mitigate first and investigate later. Afterwards write a blameless postmortem that asks "how did our system and process allow this?" rather than "who did this?". That is how reliability actually improves.

  A worked example: error budgets and burn rate

  Your service has an SLO of 99.9% of requests succeed over 30 days, and it serves 10 million requests a month.

  • Error budget: 0.1% of 10,000,000 = 10,000 failed requests may happen in the month before you break the promise.
  • Burn rate is how fast you spend it. A burn rate of 1 uses exactly the budget over 30 days. If 5% of requests are failing now, the error rate is 50 times the allowed 0.1%, a burn rate of 50: the whole month's budget is gone in 30 / 50 = 0.6 days, about 14 hours.
  • Alert on burn rate, not on a single bad minute. A common pattern watches two windows together: page someone if the burn rate over the last hour is above about 14 and the last 5 minutes agree (a fast burn: 2% of the monthly budget gone in an hour). Open a ticket for a slower burn, say rate above 6 over 6 hours. Two windows avoid paging for blips and avoid missing slow leaks.

  Why this beats "alert when error rate is above X": it ties alerts to what users were promised, it pages only when the budget is in danger, and it gives the team a rule for trade-offs: when the budget is healthy, ship faster; when it is nearly spent, slow down and fix reliability.

  100%

  Figure SDA-FND-11-2 · Alerting on how fast the error budget burns · © forcsguy.com

  Finding the cause with all three signals

  A worked investigation shows why you want metrics, traces and logs together.

  1. A metric alerts: the checkout success rate dropped and p99 latency rose. Metrics tell you that something is wrong and how bad, cheaply and over time.
  2. Narrow it down by dimension: break the metric down by region, version and endpoint. Only the new version in one region is affected, a strong hint it is a deploy.
  3. A trace shows where the time goes: one slow checkout request shows 4.8 of its 5 seconds in a call to the payment provider's tokenisation endpoint.
  4. Logs give the detail: filter logs by the trace id from that request and read the exact error: a timeout talking to a new tokenisation URL.
  5. Fix and verify: roll back the version, and watch the metric recover.

  100%

  Figure SDA-FND-11-3 · From an alert to a root cause · © forcsguy.com

  The glue is a correlation id (trace id) carried on every request and written into every log line, so you can jump from one signal to the others.

  Mistakes people make

  • Alerting on causes instead of symptoms. "CPU is 85%" may be fine; "users are seeing errors" never is. Page on symptoms users feel, and use cause metrics for diagnosis.
  • Averaging percentiles. The mean of two hosts' p99 is meaningless; merge distributions instead (see the t-digest lesson).
  • High-cardinality labels. A label such as user_id on a metric creates a new time series per user, which can multiply storage and cost by millions. Put such detail in logs or traces.
  • Logs without structure or ids. Free text is hard to query. Log JSON with a trace id, service and version.
  • Dashboards no one owns. A wall of graphs nobody reads is noise. Build a small dashboard per service around the golden signals: latency, traffic, errors and saturation.
  • No runbook. An alert that does not say what to check first wastes the first ten minutes of every incident.

  Try it yourself

  Run any test in the System Builder and read its charts as if they were production graphs: which signal moves first, and which part would you look at next? Build the search metrics lab to practise percentile maths, and read the t-digest lesson for how real systems summarise latency at scale.

  Real-world reading

  Related reading

  • The Site Reliability WorkbookGoogle SRE team · O'Reilly / GoogleBooks
  • Feature Toggles (aka Feature Flags)Pete Hodgson · Martin FowlerBlogs
  • InfluxData BlogInfluxData engineers · InfluxDataBlogs
  • MetricsDB: TimeSeries Database for storing metrics at TwitterTwitter engineers · TwitterCase Studies

  Show 6 moreShow fewer

  • The Architecture Twitter Uses to Deal with 150M Active Users, 300K QPS, a 22 MB/S Firehose, and Send Tweets in Under 5 SecondsHigh Scalability engineers · High ScalabilityCase Studies
  • Steady State: Always Put Logs on Separate DiskMicrosoft engineers · MicrosoftCase Studies
  • XCMetrics: All-in-One Tool for Tracking Xcode Build Metrics at SpotifySpotify engineers · SpotifyCase Studies
  • Honeycomb BlogHoneycomb engineers · HoneycombBlogs
  • BookKeeper: Distributed Log Storage at YahooYahoo engineers · YahooCase Studies
  • Analyzing V8 Garbage Collection Logs at AlibabaAlibaba engineers · AlibabaCase Studies

  Open this lesson with its quiz and exercise

• Messaging and Microservice Patterns · 13 minObservability: metrics, SLIs, SLOs and alertingTurning raw signals into targets people can act on: the three pillars, the four golden signals, SLIs, SLOs, error budgets and alerts that wake you only when it matters.Read hereClose

  Monitoring versus observability

  Monitoring watches known questions ("is the error rate high?"). Observability is the ability to ask new questions of a running system from the data it emits, which matters when failures are unexpected. It rests on three signal types, often called the pillars:

  Signal	What it is	Good for
  Metrics	Numbers over time (counts, rates, durations)	Dashboards, alerts, trends; cheap to store
  Logs	Timestamped event records	Detail about what happened in one place
  Traces	A request's path across services	Finding where time and errors come from

  Together with context (deploy markers, config changes) they let you move from "something is wrong" to "this is why". (Collection architecture is in the metrics monitoring lesson; instrumentation in the previous lesson.)

  What to measure

  Two handy checklists:

  • The four golden signals (for services): latency (separate successful from failed), traffic (demand), errors (rate of failures) and saturation (how full the most constrained resource is).
  • RED for request-driven services (Rate, Errors, Duration) and USE for resources (Utilisation, Saturation, Errors).

  Use percentiles (p50, p95, p99) rather than averages for latency, because averages hide the slow tail that users feel (see the latency lesson). Keep label cardinality low (do not label metrics with user ids).

  SLIs, SLOs and SLAs

  100%

  Figure SDA-MSV-06-1 · From measurement to promise · © forcsguy.com

  • SLI (indicator): a ratio of good events to valid events, such as successful requests / valid requests or requests under 300 ms / all requests.
  • SLO (objective): a target for the SLI over a window. Few, meaningful SLOs beat many.
  • SLA (agreement): the contractual promise and what happens if missed; keep it looser than the SLO so you have room to react.
  • Error budget = 1 - SLO. With 99.9 percent over 30 days you may be "down" about 43 minutes. Common reference points: 99 percent ≈ 7.2 hours a month, 99.9 percent ≈ 43 minutes, 99.99 percent ≈ 4.3 minutes.

  Using the budget: while budget remains, ship features and take risks; when it is nearly spent, prioritise reliability work. This turns arguments between "ship faster" and "be safer" into a shared, measurable policy.

  Alerting that deserves a human

  Alert fatigue is a reliability problem: people ignore noisy pages.

  • Alert on symptoms, not causes: "users see errors" rather than "one pod restarted". Keep cause metrics for dashboards and diagnosis.
  • Every page must be urgent, actionable and with a runbook; if not, make it a ticket or delete it.
  • Use burn-rate alerts: page when the error budget is being consumed much faster than allowed (for example 2 percent of a 30-day budget in an hour), ticket for slow burns. Use two windows (short and long) to be fast but not flappy.
  • Group and deduplicate related alerts; route by ownership; add severity.
  • Review alerts after incidents and prune them regularly.

  100%

  Figure SDA-MSV-06-2 · From signal to action · © forcsguy.com

  The tools, in one table

  Need	Common tools
  Metrics collection and alert rules	Prometheus (pull model), cloud metrics, Datadog
  Dashboards	Grafana, vendor dashboards
  Logs	ELK or OpenSearch stack, Loki, Splunk, cloud logging
  Traces	Jaeger, Zipkin, Tempo, vendor APM
  Instrumentation standard	OpenTelemetry
  Incident paging	PagerDuty, Opsgenie and similar

  Dashboards that help

  One overview per service showing the golden signals and SLO burn; drill-down panels for dependencies and resources; deploy and config markers on graphs; and consistent layout across services so anyone on call can read any of them.

  Incidents and learning

  Have on-call rotations with clear escalation, runbooks for known failures, and blameless post-incident reviews that produce concrete follow-ups. Practise with game days. Reliability is a process as much as a technology: measure, set targets, spend the budget wisely, and learn from every failure.

  Real-world reading

  Related reading

  • The Site Reliability WorkbookGoogle SRE team · O'Reilly / GoogleBooks
  • Google SRE Book: Service Level ObjectivesGoogle SRE team · GoogleBlogs
  • Alerting on Service-Level Objectives (SLOs) at SoundCloudSoundCloud engineers · SoundCloudCase Studies
  • PrometheusPrometheus · CNCF / PrometheusGitHub

  Show 6 moreShow fewer

  • Prometheus overviewPrometheus · CNCF / PrometheusDocumentation
  • Google SRE Book: Monitoring Distributed SystemsGoogle SRE team · GoogleBlogs
  • Grafana LabsGrafana Labs engineers · Grafana LabsBlogs
  • Monitoring and Alert System using Graphite and Cabot at HackerEarthHackerEarth engineers · HackerEarthCase Studies
  • Kotlin at DoorDashDoorDash engineers · DoorDashCase Studies
  • ThirdEye: Monitoring Platform at LinkedInLinkedIn engineers · LinkedInCase Studies

  Open this lesson with its quiz and exercise

• Databases Deep Dive · 11 minNormalisation and schema designOrganising tables so each fact is stored once, the first three normal forms in plain words, and when to break the rules on purpose.Read hereClose

  The idea in one sentence

  Normalisation means organising data so that each fact is stored in exactly one place. When a fact is stored in many places, changes must be made in many places, and sooner or later they disagree.

  The anomalies it prevents

  Imagine one big orders table with the customer's name and email on every row.

  • Update anomaly: the customer changes email; you must update every order row, and missing one leaves two different emails.
  • Insert anomaly: you cannot record a new customer until they place an order.
  • Delete anomaly: deleting a customer's only order deletes the only record of the customer.

  The first three normal forms, in plain words

  100%

  Figure SDA-DDV-01-1 · Normal forms as three questions · © forcsguy.com

  1. 1NF: no repeating groups or lists in a cell (not tags = "a,b,c"; use a related table); each row has a primary key.
  2. 2NF: relevant when the key has several columns (like order_id, product_id). A column such as product_name depends on only product_id, so move it to a products table.
  3. 3NF: a column must depend on the key, "the whole key and nothing but the key". If zip_code determines city, store city with the zip code table, not on every customer row.

  Most application schemas aim for 3NF and stop there.

  Keys and relationships

  • Primary key: unique row identity. Prefer a stable surrogate id (for example a number or UUID) over data that can change, such as an email.
  • Foreign key: a column referencing another table's key; the database then guarantees the referenced row exists.
  • One-to-many: put the foreign key on the "many" side (orders.customer_id).
  • Many-to-many: use a join table with two foreign keys (book_authors).
  • Constraints (not null, unique, check) are free correctness: let the database enforce what it can.

  CREATE TABLE customers (id BIGINT PRIMARY KEY, email TEXT NOT NULL UNIQUE, name TEXT NOT NULL);
  CREATE TABLE orders (
    id BIGINT PRIMARY KEY,
    customer_id BIGINT NOT NULL REFERENCES customers(id),
    created_at TIMESTAMPTZ NOT NULL DEFAULT now()
  );
  CREATE TABLE order_items (
    order_id BIGINT NOT NULL REFERENCES orders(id),
    product_id BIGINT NOT NULL REFERENCES products(id),
    quantity INT NOT NULL CHECK (quantity > 0),
    unit_price_cents INT NOT NULL,            -- copied on purpose: the price at purchase time
    PRIMARY KEY (order_id, product_id)
  );
  

  Notice unit_price_cents is deliberately duplicated: the order must remember what was charged even if the product price changes. That is a snapshot, not an anomaly; the distinction is whether the value should change when the source changes.

  Denormalisation: breaking the rules on purpose

  Normalised data needs joins to read. Sometimes reads dominate and joins are too costly. Common, deliberate denormalisation:

  • Precomputed counts and totals (likes_count on a post) updated with each change.
  • Copies of frequently read fields (author name on a post row) to avoid a join.
  • Summary or reporting tables rebuilt periodically.
  • Document stores that embed related data inside one record.

  The price is keeping copies in sync: update both in one transaction, rebuild from the source on a schedule, or accept eventual consistency. Rule of thumb: normalise first; denormalise a specific, measured hot path, and document it.

  OLTP and analytics

  Transactional systems (OLTP) favour normalised schemas for correct writes. Analytical systems (OLAP) favour wide, denormalised tables, often a star schema (a central fact table of events with dimension tables around it), because they scan lots of data and rarely update it.

  Habits that prevent pain

  • Model the real-world facts first, then tables.
  • Use clear, consistent names; pick one convention.
  • Store money as integers (minor units) or exact decimals, never floating point.
  • Store timestamps in UTC with time zone awareness.
  • Prefer soft deletes (a deleted_at column) only where you need history, and index accordingly.
  • Plan changes as migrations (small, reversible, tested), not manual edits.

  Real-world reading

  Related reading

  • LISA '10 - Operations at Twitter: Scaling Beyond 100 Million UsersYouTubeYouTube
  • Espresso: Distributed Document Store at LinkedInLinkedIn engineers · LinkedInCase Studies
  • Akutan: Distributed Knowledge Graph Store at eBayeBay engineers · eBayCase Studies
  • How Twitter Stores 250 Million Tweets a Day Using MySQL - High Scalability -High Scalability engineers · High ScalabilityCase Studies

  Show 6 moreShow fewer

  • TAO: Distributed Data Store for the Social Graph at FacebookFacebookPapers
  • Scaling Infrastructure at Instagram - Lisa Guo, Instagram EngineeringYouTubeYouTube
  • Cassandra for Keeping Counts In Sync at SoundCloudSoundCloud engineers · SoundCloudCase Studies
  • Roshi: Distributed Storage System for Time-Series Event at SoundCloudSoundCloud engineers · SoundCloudCase Studies
  • What Powers Instagram: Hundreds of Instances, Dozens of TechnologiesTumblr engineers · TumblrCase Studies
  • Zipkin: Distributed Systems Tracing at TwitterTwitter engineers · TwitterCase Studies

  Open this lesson with its quiz and exercise