Checks from outside our own network
Monitors in Dublin, Ohio and Sydney call each service the way a customer would. A component turns amber only when two of the three agree, which keeps one bad route from raising a false alarm.
All systems operational
Halyard Cloud is running normally in every region.
This page shows the health of every Halyard Cloud service as our monitors see it, checked every thirty seconds from three continents. When something breaks, the first post goes up within ten minutes, and it says what is affected, what we know and when the next update will arrive.
Right now
- Compute and containers
- Operational
- Managed databases
- Operational
- Object storage
- Operational
- Dashboard and API
- Operational
Last ninety days
- Dashboard and API uptime
- 99.99%
- Compute uptime across all regions
- 99.97%
- Incidents with customer impact
- 2
- Median time to the first public update
- 11 min
Uptime by component and region
Minutes of customer-visible downtime over the last ninety days, measured by synthetic checks that run every thirty seconds.
| Compute and containers | Frankfurt | 26 | 41 |
|---|---|---|---|
| Compute and containers | Virginia | 39 | 38 |
| Compute and containers | Singapore | 52 | 52 |
| Managed databases | Frankfurt | 13 | 6 |
| Managed databases | Virginia | 26 | 5 |
| Object storage | Frankfurt | 0 | 64 |
| Object storage | Virginia | 13 | 59 |
| Dashboard and API | Global | 13 | 120 |
How we keep this page honest
An engineer is always on call
One primary and one backup engineer carry the pager every week, on a rotation that follows the sun. If the primary does not acknowledge within five minutes, the backup is paged.
Fixes are done in the open
Customers see which step the team is on: investigating, identified, fixing or monitoring. Each state change is time-stamped, so you can plan around it instead of refreshing.
Updates reach you first
Subscribe by email or SMS to the services and regions you use. A message arrives when an incident starts, whenever it changes and once more when it is resolved.
Past incidents
Every incident with customer impact in the last six months, with the cause in plain words.
- September 2026
Slow image pulls in Singapore · Compute and containers · 38 minutes
A full disk on one cache node slowed container starts for about a fifth of Singapore customers. We drained the node, added disk alarms to every cache and found no data loss.
- July 2026
Dashboard sign-in errors · Dashboard and API · 21 minutes
A certificate on an identity service expired before its renewal job ran. Sign-ins failed for 21 minutes; running services were not touched. Renewals now alert 14 days ahead.
- May 2026
Delayed database backups in Virginia · Managed databases · 3 hours
Nightly backups ran three hours late after a storage upgrade. Every backup completed and passed its restore test, and no database was unavailable.
- April 2026
Object storage upload errors · Object storage · 17 minutes
A change to request limits rejected some multipart uploads in Frankfurt. We reverted the change and added a check that replays large uploads before every release.
Get alerts for the services you use
Enter your email and we will send a message when an incident begins, changes or ends. You can narrow it to a single region from the first email.
About this page
What do the colors mean?
Green means operational, amber means degraded performance or a partial outage and red means a major outage. Blue marks scheduled maintenance that we announced at least five days ahead.
How do I report a problem you have not posted?
Write to support from your dashboard with your region and a request ID. If several customers report the same thing, we open an incident even before our monitors agree.
Do you offer service credits?
Yes. If monthly uptime for a paid service falls below the commitment in your plan, the credit is applied to your next invoice without you having to ask.
Can I put this status on my own site?
A JSON feed and an embeddable badge show the current state of any component, so you can display it inside your own product.