Coming from PagerDuty
Tissue Mast has an importer that reads your PagerDuty account and builds the same rotations, escalation policies and alert sources here, with a dry run first that changes nothing. It's in the web app under On call, as Import from PagerDuty, and it's open on every plan. The importer below says how it reads each part. What it can't move is the senders: PagerDuty's integration keys stay in PagerDuty, so each sender gets repointed at a send URL.
The rest of this page is the word list: what PagerDuty calls something, and what it's called here.
The words
| PagerDuty | Tissue Mast | Notes |
|---|---|---|
| Service with an integration key | Alert source | One alert source does what a service and its integration do together: it has the send URL, the default level, the highest level and the escalation policy. The API calls it a channel |
| Integration key | Send key (mk_…) |
Mast doesn't speak the Events API. Senders post to the send URL instead. Sending alerts |
| Incident | Incident | The same idea: one thing that went wrong, from the first alert to the recovery |
| Urgency high and low, priority P1 to P5 | Level: page, loud, normal, quiet | High urgency is closest to page. Sends that say P1, critical or high are translated. Priority names from other tools |
| Static low urgency on a service | Highest level | The nearest thing, not the same thing. A highest level is a cap: a send asking for more goes out at the highest level and the answer says "capped": true. Only a phone can raise it to page |
| Alert grouping | Grouping | Mast's grouping collapses a flood on one alert source into one card with a count. A page is never grouped away |
| Escalation log, incident timeline | Timeline | |
| Schedule, layer | Rotation | Each layer becomes its own rotation. Mast doesn't stack layers, so where a higher layer covered a lower one, both are on call |
| Override | Override | |
| Escalation policy | Escalation policy | Steps page a rotation, the next person on it, one person, everyone or another alert source |
| Support hours | Mute hours | Close, not the same. Mute hours are a daily window on one alert source when its alerts arrive without a sound. A page still rings |
| Contact methods, notification rules | Devices | Mast pages a linked phone or watch through push, and calls from a step with Call ticked. There are no per-person rules and no SMS |
| Maintenance window | Maintenance window | |
| Acknowledge, resolve | Acknowledge, resolve | |
| (no first-class object) | Vital (heartbeat) | A URL a job beats on a schedule. Mast pages when the beats stop |
Where it isn't a match
Event rules and routing. Mast has no rules that send one kind of alert to one policy and the rest to another. Make two alert sources, each with its own policy, and point each kind of alert at its own send URL.
New alert sources top out at loud. One made by the importer, in the web app or from the command line can't page until someone raises its highest level from a phone. Do that for the ones that should wake people, then send a test page to check.
Business services and impact. There's no service graph. Status pages show what your probes and vitals see.
Round robin. An escalation rule that hands its targets out in turn has no equivalent. Mast pages every target of a step at once, and the dry run names each rule that did it.
The importer
You need to be the account's owner or an admin. In PagerDuty, make a REST API key under Integrations › API Access Keys. A read-only key is enough, and it's all the importer needs. Paste it on the import screen, pick the region, US or EU (EU if your PagerDuty address ends in eu.pagerduty.com), and press Read PagerDuty. Mast reads your configuration through PagerDuty's REST API, uses the key for that request only and keeps no copy, not in its records, its logs or the account's activity. Delete the key in PagerDuty when you're done.
The first read is a dry run and changes nothing. It's open on every plan, Free included, so you can see what a move involves before you pick one. It lists every rotation, escalation policy and alert source Mast would make, and for each one whether it's new, changes something that's already there, is already the same, is left out, or doesn't fit the plan. Above the lists it sums up the headcount, the teams the key can see, how many people PagerDuty texts or phones, and the smallest plan that holds the move. The Opsgenie importer works the same way and has the long version.
Untick anything you don't want, then press Import the items ticked. Mast reads PagerDuty again at that moment, so a change made there in between is picked up. Everything is matched by name, so running it again changes what's there instead of making copies, and finishes the job if something failed partway.
People are matched to members of this account by email address, ignoring case. The importer never invites anyone on its own: everyone it can't match gets a tick box, and Invite sends each person ticked a normal member invite. Someone who isn't on the account yet is left out of their rotations and escalation steps until they've joined and you run the import again.
How each part maps
- Teams. When the key sees more than one team, a schedule, policy or service that belongs to a team gets the team's name in front: the Platform team's "Primary" schedule becomes
Platform / Primary. With one team or none, names stay as they are. - Schedules. Each layer becomes its own rotation. A schedule with one layer keeps the schedule's name, and one with several becomes
<schedule> / <layer>for each. A daily turn stays daily, a weekly one weekly and a two-week one every two weeks; any other length in whole hours becomes a custom length, up to 720 hours. Turns are counted from the layer's own start, so nobody's turn moves on the day you switch. A restriction, such as weekdays from 09:00 to 17:00, is dropped, so the rotation is on around the clock, and the dry run says what the restriction was. - Overrides. Overrides in the next 90 days come across as timed overrides, on every rotation made from that schedule.
- Escalation policies. Each rule becomes a step. PagerDuty counts each rule's delay from the rule before it, and Mast counts each step's wait from the alert, so the delays are added up: rules 30 minutes apart become steps at 0, 30 and 60 minutes. A rule that names a schedule pages whoever is on call there, and a rule that names a person pages that person. "Repeat this policy" carries over as going round again, the same number of times.
- Services. Each service becomes an alert source on its escalation policy, with the service's integrations named in a note so you know which senders to repoint. A disabled service is made switched off. A service set to constant low urgency gets a note, since low urgency never paged in PagerDuty and a Mast alert source can. Alert grouping, auto-pause, support hours and the acknowledgement and auto-resolve timeouts are listed as not imported.
When the import finishes, it prints every alert source with its send URL. The key in the URL is the whole secret, so treat the list like a password. It also gives you a Terraform import block for everything it made or changed, as the Opsgenie importer does.
Mast has nowhere to put these, so the importer leaves them in PagerDuty and lists them once: personal notification rules, on-call handoff notifications, event orchestration, rulesets and suppression, response plays and incident workflows, business services and service dependencies, status pages, extensions and webhooks, incidents and incident history.
The importer reads users, teams, schedules, escalation policies and services, and nothing else. From each person it keeps the name, the email address and the kind of contact method they have, never a phone number. From each service it keeps the integration names, never their keys or addresses.