Quick switcher (Go to…)
Last updated 13 November 2026
Press ⌘P (Ctrl+P on other systems), or click Go to… in the title bar, to jump to a page or to one resource on it: a Lambda function, a log group, a queue, an alarm. It works from anywhere in the app, even while you are typing in a field or editor, but not over another open dialog.
What it searches
- Pages: every service on your rail, by name and summary.
- Resources ARNscope has already listed in the current profile and region: lists the app loaded at start-up or when you opened a page. Opening the switcher takes a snapshot of those lists, and every keystroke searches only that snapshot.
The switcher never calls AWS. A resource whose list isn’t loaded yet can’t be found until you open its page once (see Why a resource is missing).
| Service | Found by |
|---|---|
| AWS Lambda | Name, ARN, log group, tags, runtime |
| Amazon ECS | Clusters by name and ARN; services (name, ARN, cluster) only for clusters you have opened |
| Amazon EC2 | Name, instance ID, private and public IP |
| AWS Step Functions | Name, ARN |
| Amazon RDS | Identifier, ARN, endpoints, engine; marked Cluster or Instance |
| Amazon DynamoDB | Table name (the first page of tables) |
| Amazon SQS | Name, queue URL |
| Amazon MSK | Name, ARN |
| Valkey | Cache ID, ARN, primary and reader endpoints, engine |
| Amazon S3 | Bucket name, s3://bucket, bucket ARN |
| Load balancer access logs | Name, ARN, DNS name |
| VPC flow logs | Name, VPC ID, CIDR blocks |
| CloudWatch Logs | Log group name, ARN |
| CloudWatch dashboards | Dashboard name (dashboards in AWS) |
| CloudWatch alarms | Name, ARN, metric name, namespace |
| StarRocks, ARNscope tunnels, SOPS secrets | The page only |
ARNs, URLs and endpoints match by their own part (an ARN’s resource, a URL’s last segment, a host’s name before the region), so typing aws, east or a region doesn’t list every resource. A whole ARN, URL or host matches when the token has a :, / or . in it. A 12-digit account ID matches resources in that account exactly. A log group’s leading /aws/ is ignored.
How matching works
- Several words narrow: every word must match (AND), in any field.
prod checkoutfinds items with both. - Best match first: an exact word, then the start of a word, then text inside a word, then a typo, then scattered letters. Pages rank a little above resources, and items you opened recently a little higher still.
- Fuzzy: a word of 5 or more letters can match scattered letters in a name, as long as it starts at a word and most letters are together or at word starts:
ckapifindscheckout-api. - Typos: a word of 5 or more letters with one letter wrong, missing, extra or swapped finds the name (
paymnet→payment-capture), when nothing matches it as typed. The first letter must be right. - Fuzzy and typo matches apply to names only, never to ARNs, tags or service types. Service words (
ecs,queue,postgres) are never matched by scattered letters. - Short words: a one- or two-letter word matches a resource’s service type only when it is the whole type (
s3), soeclists pages, not every EC2 instance. - Engines and runtimes:
postgres,mysql,mariadb,aurora,redisorvalkeyfind databases and caches of that engine;pythonornodejsfind Lambda functions on that runtime. - Service words: a service’s name, ID or alias (
function,queue,bucket,db,kafka,cache…) matches everything of that type:ecs checkoutputs ECS resources named checkout first.
Results are grouped by service, at most 50 at a time. When there are more, the footer says Showing 50 of N. Type more to narrow. When two or more pages match, they lead in a Pages group. A row shows why it matched when the name doesn’t say (the tag or ID that matched, highlighted), and rows with the same name show their ID.
Qualifiers
| Qualifier | Searches | Examples |
|---|---|---|
tag: |
Tags as key=value: the key, the value or both |
tag:env=prod, tag:prod, tag:team |
type: |
The service’s name, ID or an alias | type:lambda, type:queue, type:db |
id: |
IDs, ARNs, URLs, endpoints and the account ID | id:i-0abc, id:arn:aws:sqs, id:123456789012 |
<service>: |
Only that service; text after the colon is matched too | ecs:, rds:orders, lambda:checkout |
The service prefix takes a service ID (alarms, cloudwatch, logs, lambda, ecs, ec2, sfn, rds, dynamodb, sqs, msk, valkey, starrocks, s3, alb, vpc, tunnels, secrets) or a one-word alias: alarm, dashboard, serverless, vm, database, db, dynamo, ddb, queue, kafka, elasticache, olap, bucket, elb, tunnel, ssm, sops.
Words that also appear inside ARNs (cluster, instance, container, function, table, cache, secret, network, server, storage, workflow, metrics) are not qualifiers, so a pasted ARN fragment like cluster:commerce is searched as text. URLs (s3://…, https://…) and ARNs are never read as qualifiers.
There is no svc: qualifier: use the service itself as the prefix (lambda:). An unknown prefix that matches nothing shows Unknown qualifier svc:.
Tags: Lambda functions carry all their tags. EC2 instances and VPCs carry only their Name tag. Other resources have no tags in their lists.
Paste an S3 location
Paste an s3://bucket/key URI and the first row, S3 location, opens that bucket at that key. The key is taken literally (spaces, %, ?, #, * included).
Recently opened
With nothing typed, the switcher lists Recently opened, then every page.
- Recents are kept per profile, on this Mac: up to 20, of which at most 3 are pages, so visits from the rail don’t push out resources.
- Pages you open from the rail or a link count too.
- A recent resource shows only in the region you opened it in. It stays findable even before its list is loaded again.
- When the first recent item is where you already are, the next one is preselected, so ↵ takes you somewhere new.
Keys
| Key | Does |
|---|---|
| ⌘P / Ctrl+P | Open; pressed again while open, move down |
| ↑ ↓ | Move through the results (wraps around) |
| ⌘↑ ⌘↓ (Ctrl+Home / Ctrl+End elsewhere) | First or last result |
| Page Up Page Down | A screenful up or down |
| ↵ | Open the selected page or resource |
| Esc | Close and go back to where you were |
Home and End stay in the search box, and Tab doesn’t leave it. The switcher also closes when the ARNscope window loses focus. The footer shows the profile and region it searches (profile default when the profile’s own region is used). After opening a resource, focus moves to its row in the page’s list.
Why a resource is missing
| Cause | What to do |
|---|---|
| Its list isn’t loaded yet. The footer says so for a service you type, for example Amazon SQS: searchable once its page is opened, or Some lists are searchable only after their page is opened when nothing matches. | Open the service’s page once. Its list is then cached and searchable. |
| It is an ECS service in a cluster you haven’t opened (Amazon ECS services: searchable once their cluster is opened). | Open the cluster on the ECS page. |
| It is in another profile or region. | Switch the profile or region first; the switcher searches the current ones only. |
| Its service is hidden from the rail. | Show the service again; hidden services aren’t searched. |
| It is newer than the cached list. | Refresh the service’s page. |
| Its service is listed as pages only (StarRocks, tunnels, SOPS secrets). | Open the page and pick it there. |
Limitations
- It searches only lists ARNscope has already loaded, by design: it never calls AWS, so it costs nothing and never slows down.
- DynamoDB tables come from the first page of the table list; ECS services only from clusters you have opened.
- Your own CloudWatch dashboards (My dashboards) aren’t searched; dashboards in AWS are.
- Tags are searchable for Lambda functions, and the
Nametag for EC2 instances and VPCs, only. - StarRocks clusters, tunnels and SOPS secrets are reachable as pages only.
Want something else to be searchable? Write to [email protected] and tell us what you look for.