Slack's Migration to a Cellular Architecture
Comments on “Slack's Migration to a Cellular Architecture”
Claim-checked. Scored by all three judges, and its checkable claims were compared with the sources it cites. How far stories are checked
Slack's engineering blog describes moving its core user-facing services to a cellular architecture over about 1.5 years, prompted by a June 2021 us-east-1 network fault that users still noticed despite multi-AZ hosting. Each availability zone is now a silo, with services talking only within their zone. An AZ drain then works by reweighting Envoy edge load balancers through its Rotor xDS control plane, targeting 1% granularity and propagation in seconds. These figures and the timeline come from Slack itself and were not independently verified.
Why it ranked
No. 7 of 60 with a score of 0.70. The panel rated it highest on new information (0.92) and technical depth (0.80), lowest on significance (0.61).
- AI judgeJudge 2claude-haiku-5-5
- Detailed engineering write-up of a cellular AZ-drain architecture with Envoy weighted clusters and siloing; concrete, practically useful for infrastructure builders, with reasonable but company-internal evidence.
- AI judgeJudge 3gpt-5.6-luna
- A detailed, evidence-backed engineering account of migrating critical services to cellular, siloed architecture, including failure analysis, operational goals, Envoy/xDS implementation, and trade-offs. Highly actionable for reliability and infrastructure practitioners.
Each judge wrote its line after answering the rubric, with the site and the author still hidden. Every answer behind this is in the table under Score.
Score
0.70total under rubric v2.3
- Significance0.61
- Reader interest0.78
- Audience breadth0.63
- Practical relevance0.76
- Substantiated0.77
- Technical depth0.80
- New information0.92
- Counts against: Promotional0.08
- Counts against: Overclaim0.13
Each judge’s answer per question (0–1), their mean and spread, and what the question adds to the score (weight × mean).
| Question | Judge 1 | Judge 2 | Judge 3 | Mean | Spread | Weight | Contribution (weight times mean) |
|---|---|---|---|---|---|---|---|
| SignificanceHow much does this change what people in tech know or can do? | 0.61 | 0.45 | 0.78 | 0.61 | 0.33 | +0.24 | +0.147 |
| Reader interestWould a well-informed senior engineer or tech founder want to read this today, and likely pass it on? | 0.65 | 0.80 | 0.90 | 0.78 | 0.25 | +0.26 | +0.204 |
| Audience breadthHow many people who build technology does it concern? | 0.51 | 0.67 | 0.72 | 0.63 | 0.21 | +0.13 | +0.082 |
| Practical relevanceHow useful is it to someone who builds software or tech products? | 0.72 | 0.70 | 0.86 | 0.76 | 0.16 | +0.22 | +0.168 |
| SubstantiatedHow well is it backed? Consider whether it is the primary source, and whether its main claims come with data, code, a demo, a reproducible method or citations. | 0.66 | 0.75 | 0.90 | 0.77 | 0.24 | +0.05 | +0.038 |
| Technical depthHow technically substantive is it? | 0.78 | 0.80 | 0.83 | 0.80 | 0.05 | +0.05 | +0.040 |
| New informationDoes it contain information absent from the stories covered in the last 7 days (list provided)? | 0.89 | 0.90 | 0.96 | 0.92 | 0.07 | +0.05 | +0.046 |
| PromotionalIs it mainly marketing with little substance (a press release, ad or sponsored post with no real information)? A launch post that carries real news is not promotional. | 0.07 | 0.10 | 0.08 | 0.08 | 0.03 | −0.19 | −0.016 |
| OverclaimDo the headline or conclusions claim more than the body supports? | 0.17 | 0.15 | 0.06 | 0.13 | 0.11 | −0.11 | −0.014 |
| Sum of contributions | +0.696 | ||||||
| Claim-check penalty | 0.000 | ||||||
| Total, as published | 0.696 | ||||||
Spread is the highest minus the lowest judge answer. marks a spread above 0.40: the judges disagree, and the story is marked contested.
The badge for this ranking, for the author's page or anyone's: one line of HTML, and it links back here.
<a href="https://rubric.hrih.si/editions/2026-11-16/7/"><img src="https://rubric.hrih.si/badge/2026-11-16/7.svg" alt="Ranked #7 on Rubric, 16 Nov 2026" width="238" height="28"></a>Claim check
Claim check: 6 no source given · 2 couldn’t check. Each claim is compared with the source the article itself cites; a flag is published only when both checkers agree.
No source givenNumber
Slack migrated its most critical user-facing services from a monolithic to a cell-based architecture over the last 1.5 years.
The article gives no link or citation for it.
The article says “we have migrated the most critical user-facing services at Slack from a monolithic to a cell-based architecture over the last 1.5 years”
The source says No source to compare with.
No source givenDate
On 2021-06-30 the cloud provider's network link in one availability zone in us-east-1 began intermittent faults at 11:45am PDT, was removed automatically at 12:33pm, and faulted again at 5:22pm before being permanently removed at 5:31pm PDT.
The article gives no link or citation for it.
The article says “At 5:22pm PDT on 2021-06-30, the same network link experienced the same intermittent faults. At 5:31pm PDT on 2021-06-30, the cloud provider permanently removed the network link from service”
The source says No source to compare with.
Couldn’t checkNumber
Slack's 99.99% availability SLA allows less than 1 hour per year of total unavailability.
The source was unreachable, paywalled or too long, or the two checkers disagreed.
The article says “Slack’s 99.99% availability SLA allows us less than 1 hour per year of total unavailability”
The source says No passage could be compared.
Source cited by the article: slack.com/terms/service-level-agreement
No source givenNumber
A single Slack API request, such as loading messages in a channel, may fan out into hundreds of RPCs to service backends.
The article gives no link or citation for it.
The article says “A single Slack API request from a user (for example, loading messages in a channel) may fan out into hundreds of RPCs to service backends”
The source says No source to compare with.
No source givenNumber
Slack's design goal was to remove as much traffic as possible from an AZ within 5 minutes.
The article gives no link or citation for it.
The article says “Remove as much traffic as possible from an AZ within 5 minutes”
The source says No source to compare with.
Couldn’t checkVersion
Slack's edge load balancers all run Envoy and receive configuration from Rotor, its in-house xDS control plane, after migrating from HAProxy.
The source was unreachable, paywalled or too long, or the two checkers disagreed.
The article says “all our edge load balancers are now running Envoy and receive configuration from Rotor, our in-house xDS control plane”
The source says No passage could be compared.
Source cited by the article: slack.engineering/migrating-millions-of-concurrent-websockets-to-envoy
No source givenCausal conclusion
Siloing means each service only sends and receives traffic within its AZ, so failure in one AZ is contained and traffic can be routed away at the frontend.
The article gives no link or citation for it.
The article says “Failure of a system within one AZ is contained within that AZ, and we may dynamically route traffic away to avoid those failures simply by redirecting at the frontend”
The source says No source to compare with.
No source givenNumber
Control plane propagation of drain weights takes on the order of seconds and weights allow drains with 1% granularity.
The article gives no link or citation for it.
The article says “Propagation through the control plane is on the order of seconds; Envoy load balancers will apply new weights immediately”
The source says No source to compare with.
Also covered by
No other coverage was clustered with this story.
Elsewhere
Elsewhere: HN 3 pts, not on front page
Recorded for context only, checked 16 Nov 2026, 05:47 UTC. It played no part in the ranking.