Create file systems (with a mount target in each default subnet unless asked
not to), add and remove mount targets, create and delete access points, and
attach a file system to a Lambda function through an access point. A function
not yet in a VPC is put in the file system's mount target subnets with the
default security group, as Lambda requires. TheCloud's mounts are real, so two
functions on one access point share files.
Deleting a file system removes its access points and mount targets first, and
is refused while a function still mounts it.
Create groups from a launch template (in the default subnets unless given),
change their minimum, desired and maximum capacity, replace or remove an
instance, and follow the launches and terminations TheCloud's reconciler makes.
Deleting a group terminates its instances. Scaling policies are not offered:
the runtime stores them but its alarms only drive ECS scaling.
Create pipes from SQS, Kinesis, DynamoDB streams or MSK to Lambda, SQS, SNS,
EventBridge buses or Step Functions, with an optional filter, enrichment and
input template; start and stop them, see the flow with links to each resource,
and send a test message to an SQS source. Sources and targets the runtime's
poller cannot deliver between are refused on create.
Request public certificates (optionally exportable) or import your own, see
each domain's DNS validation records and what uses a certificate, download the
certificate and chain, export an exportable one with its encrypted key, and
revoke it. ListCertificates asks for every key type, since AWS otherwise
leaves EC certificates out.
Console Home counted every available service through the generic resource
list, but MSK and MSK Connect are available through their own pre-SPI pages
with no adapter behind /resources, so each answered 501 and the Resources tile
read "Resource load failed". Service descriptors now say whether the generic
list serves them (listable, true when an adapter is registered), and Console
Home counts only those; the MSK cards still link to their pages.
API Gateway v2 is a separate AWS API from REST APIs, so HTTP and WebSocket APIs
made with the SDK, SAM, CDK or Terraform never showed up. The API Gateway page
now lists all three kinds with a Type column, creates each from the same form,
and opens HTTP and WebSocket APIs in their own panel:
- Quick create: an HTTP API with a Lambda gets a catch-all route on an
auto-deployed $default stage, plus the Lambda permission.
- Routes to Lambda (payload 2.0) or HTTP backends; deleting a route removes
the integration and Lambda permission only it used.
- Stages with or without automatic deployment, and deploy on demand.
- JWT authorizers filled in from a Cognito pool's OpenID Connect discovery
document and an app client, or from any issuer; protect routes with them.
- CORS configuration.
- A tester that sends a real request to a stage, and for WebSocket APIs one
that connects, sends a message and collects the replies.
Capabilities are advertised under a new httpApiActions group, covered by the
cloudProxy guard test. Verified end to end against TheCloud's runtime.
Labels, empty states, error messages and defaults the console shows now name
TheCloud or its runtimes. Identifiers that must match the emulator (FLOCI_*
settings, the X-Floci-CSRF header, emulator-returned host names) are unchanged.
Distributions were linked under localhost.floci.io, another project's domain.
Every .localhost name resolves to the machine itself, so the brand-neutral
<id>.cloudfront.localhost reaches the runtime with no DNS or hosts entry.
CI failed on a different set of tests each run because bun runs every test
file in one process and the file order differs between machines:
- the secretsmanager route test replaced the whole ../aws module, so files
loaded after it lost awsClients, lambda and defaultAccountId (API Gateway,
SES)
- the Lightsail tests restored FLOCI_ENDPOINT by assigning undefined, which
stores the string "undefined" (runtime diagnostics)
- the auth tests used full-strength argon2 for ~30 sign-ins and timed out on
slower runners; the OIDC cleanup test cached a discovery document without
userinfo_endpoint for the shared issuer
Checked with 20 runs under bun test --randomize.
The runtime leaves EstimatedNumberOfUsers at 0, so the pool metadata now
carries a userCount from ListUsers (capped at 1000), which the delete
confirmation reads.
Adds enable, disable, confirm, sign-out, password reset and set-password
for one user; user group membership; group list, create and delete; and a
test sign-in that returns decoded token claims. A refused test sign-in is
an outcome rather than a 401, which would end the console session.
Also adds a generic GET for one flat item and a schema deleteImpact so
delete confirmations can say what else goes with a resource.
Adds a userpools catalog entry, schema and AwsUserPoolsAdapter, verified
against Floci core:
- list/get/create/delete user pools; create can set email or username
sign-in, a relaxed password policy for local test users, and a public
app client (on by default) whose id is surfaced in the pool metadata
- users as the pool's child items: list (paged), get, create and delete;
passwords are set permanent by default so users can sign in at once
- map Cognito's HTTP-400 error names (UserNotFoundException,
UsernameExistsException, AliasExistsException, InvalidPasswordException)
- link Floci's cognito-idp service to the new page on the diagnostics page
The login page stored its theme under thecloud-theme while the console read floci-theme, so a choice made at sign-in was lost. Both now use floci-theme (falling back to the old key once), and useTheme tolerates unavailable storage.
Adds GET /api/clouds/:cloud/diagnostics and a Diagnostics page in the
General nav. The report combines, in one call:
- runtime reachability, latency, version and edition from /_floci/health
- every service the runtime emulates, with the console pages backed by it
- per-console-service probes (the existing status check)
- EC2 instance containers from /_floci/diagnose
- console uptime, sign-in mode and active session count
Each source is optional, so a down runtime still yields a partial report.
- CodeDeploy: query the redesigned form's labels.
- CloudFormation: expect the parameters and capabilities fields the form
now submits.
- ALB listeners: advance the five-second minimum create state before
deleting, since row actions are disabled while it is pending.
- Lightsail: cast the fetch mock as a whole so it type-checks.
- TargetGroupPanel: depend on the stable refetch function to satisfy
react-hooks/exhaustive-deps.
- safeReturnTo resolves the target against a placeholder origin and rejects
backslashes and control characters, so `/\evil.example` and similar
payloads can no longer redirect off-site after sign-in.
- Add a real unauthenticated /api/health and answer unknown /api/* paths
with a JSON 404 instead of the SPA shell.
- Limit request bodies: 16 KB for /api/auth/* (read before a session
exists), 128 MB for resource create/update and uploads, 10 MB elsewhere.
- Run the production image as an unprivileged user and add a Docker
HEALTHCHECK against /api/health.
Adds the cyber theme to the console and login page, replaces the dark/light toggle with a theme switcher, and moves the login page script to /login.js so it is served under the existing CSP.
- Key login throttling on the real peer address; read X-Forwarded-For only
from FLOCI_UI_TRUSTED_PROXIES (right-most untrusted hop). Add a per-IP
limit across usernames so one attacker can no longer lock out the admin.
- Require an OIDC allow-list (FLOCI_UI_OIDC_ALLOWED_GROUPS / _EMAILS, or
FLOCI_UI_OIDC_ALLOW_ALL_USERS=true); fall back to UserInfo for groups
when the ID token omits them. Redirect IdP and token errors to /login
instead of returning 500s; cache discovery and JWKS.
- Sweep expired sessions, attempts and pending OIDC requests every 60s,
cap each map, and throttle OIDC sign-in starts per IP.
- Fix the .ts import extension that broke the API type-check.
- Add auth.test.ts and document the new settings.
Zakaria
Zakaria