villebro commented on code in PR #233:
URL: 
https://github.com/apache/superset-kubernetes-operator/pull/233#discussion_r3650587629


##########
internal/controller/redact.go:
##########
@@ -0,0 +1,69 @@
+/*
+Licensed to the Apache Software Foundation (ASF) under one
+or more contributor license agreements.  See the NOTICE file
+distributed with this work for additional information
+regarding copyright ownership.  The ASF licenses this file
+to you under the Apache License, Version 2.0 (the
+"License"); you may not use this file except in compliance
+with the License.  You may obtain a copy of the License at
+
+    http://www.apache.org/licenses/LICENSE-2.0
+
+Unless required by applicable law or agreed to in writing, software
+distributed under the License is distributed on an "AS IS" BASIS,
+WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+See the License for the specific language governing permissions and
+limitations under the License.
+*/
+
+package controller
+
+import "regexp"
+
+// redactedPlaceholder replaces credential values matched by the redaction
+// patterns below.
+const redactedPlaceholder = "***"
+
+var (
+       // uriUserinfoRe matches the password component of a URI userinfo 
section
+       // (scheme://user:password@host). Task failure output commonly embeds 
full
+       // database connection URIs in driver error messages, e.g.
+       // "connection to postgresql://superset:hunter2@db:5432/superset 
failed".
+       // The user component may be empty (redis://:password@host). The 
password
+       // match is greedy up to the last @ in the whitespace-delimited token, 
so a
+       // malformed password containing a raw @ is masked in full rather than
+       // leaking its tail after the first @.
+       uriUserinfoRe = 
regexp.MustCompile(`([a-zA-Z][a-zA-Z0-9+.-]*://[^/@\s:]*):([^\s]+)@`)
+
+       // authorizationHeaderRe matches HTTP Authorization header values, with 
or
+       // without an auth scheme: "Authorization: Bearer eyJ...", 
"authorization=
+       // Basic dXNlcg==". The scheme word is kept; only the credential is 
masked.
+       authorizationHeaderRe = 
regexp.MustCompile(`(?i)\b(authorization\s*[=:]\s*(?:(?:bearer|basic|token|digest|negotiate)\s+)?)(\S+)`)
+
+       // bareBearerRe matches a bearer token without an Authorization prefix
+       // ("bearer eyJhbGci..."). The token must contain at least one digit so
+       // prose like "bearer authentication failed" is not mangled; real tokens
+       // (JWTs, base64, hex) virtually always contain digits.
+       bareBearerRe = 
regexp.MustCompile(`(?i)\b(bearer\s+)([A-Za-z0-9._~+/=-]*[0-9][A-Za-z0-9._~+/=-]*)`)
+
+       // credentialAssignmentRe matches free-form credential assignments such 
as
+       // "password=hunter2", "PGPASSWORD: hunter2", or "secret_key = hunter2".
+       // The key may carry prefixes/suffixes (DB_PASSWORD, api-key-prod).
+       credentialAssignmentRe = 
regexp.MustCompile(`(?i)([A-Za-z0-9_-]*(?:password|passwd|pwd|secret|token|api[_-]?key|credential|passphrase)[A-Za-z0-9_-]*\s*[=:]\s*)(\S+)`)

Review Comment:
   Before hand-rolling regex redaction, could you evaluate the existing Go 
libraries in this space and note the outcome in the PR description? The two 
closest are [`masq`](https://github.com/m-mizutani/masq) and 
[`logfusc`](https://github.com/AngusGMorrison/logfusc).
   
   My read is that neither quite fits, but it's worth stating rather than 
implying:
   
   - `logfusc` wraps secrets the program holds as typed values (`Secret[T]`) so 
they redact on log/marshal. Here the credentials arrive inline in opaque 
task-container stdout — we never hold them as typed values, so there's nothing 
to wrap.
   - `masq` is a `slog` `ReplaceAttr` filter. This redaction targets CRD 
`status` and Kubernetes Events, not a log sink, and its pattern matching would 
still need the same regexes supplied via `WithRegex` — so it adds a dependency 
and a slog impedance mismatch without removing the regexes.
   
   If that's the conclusion, a sentence or two in the PR description 
("evaluated masq/logfusc; both target secrets held as typed values or slog 
attributes, whereas we scrub unknown credentials out of opaque subprocess 
output written to status/Events") would preempt exactly this question from 
other reviewers. If one of them *does* cover it, even better — we avoid 
maintaining regexes.
   
   Worth also noting the patterns here are inspired-by / reducible-to 
[gitleaks](https://github.com/gitleaks/gitleaks)' ruleset if you want a 
reference catalog, though importing gitleaks itself would be far too heavy for 
scrubbing a 256-char string.



##########
docs/reference/releases.md:
##########
@@ -30,6 +30,10 @@ This page tracks notable changes in Apache Superset 
Kubernetes Operator releases
 - **Helm extra manifests.** The Helm chart now supports `extraManifests` for 
rendering trusted, release-scoped Kubernetes manifests with Helm `tpl`. Use it 
for companion resources owned by the operator release, not shared cluster 
infrastructure such as Gateway API controllers, CRDs, or shared Gateways 
([#196](https://github.com/apache/superset-kubernetes-operator/pull/196), 
[@younsl](https://github.com/younsl)).
 - **Helm resize policy.** The Helm chart now supports `resizePolicy` on the 
manager container, controlling whether an in-place pod resize 
(InPlacePodVerticalScaling) restarts the container 
([#200](https://github.com/apache/superset-kubernetes-operator/pull/200), 
[@younsl](https://github.com/younsl)).
 
+### Security

Review Comment:
   `### Security` is a valid [Keep a Changelog](https://keepachangelog.com) 
category, so the header is fine — but the canonical ordering places it **last** 
(Added → Changed → Deprecated → Removed → Fixed → Security). Here it's inserted 
between `### Added` and `### Changed`. Could you move the section below `### 
Changed`?
   
   (Minor: the released `0.1.0` section already carries a "Task failure 
messages may include credential fragments" entry under **Known limitations** — 
that stays as historical record; this Security entry is effectively its 
mitigation.)



##########
internal/controller/redact.go:
##########
@@ -0,0 +1,69 @@
+/*
+Licensed to the Apache Software Foundation (ASF) under one
+or more contributor license agreements.  See the NOTICE file
+distributed with this work for additional information
+regarding copyright ownership.  The ASF licenses this file
+to you under the Apache License, Version 2.0 (the
+"License"); you may not use this file except in compliance
+with the License.  You may obtain a copy of the License at
+
+    http://www.apache.org/licenses/LICENSE-2.0
+
+Unless required by applicable law or agreed to in writing, software
+distributed under the License is distributed on an "AS IS" BASIS,
+WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+See the License for the specific language governing permissions and
+limitations under the License.
+*/
+
+package controller
+
+import "regexp"
+
+// redactedPlaceholder replaces credential values matched by the redaction
+// patterns below.
+const redactedPlaceholder = "***"
+
+var (
+       // uriUserinfoRe matches the password component of a URI userinfo 
section
+       // (scheme://user:password@host). Task failure output commonly embeds 
full
+       // database connection URIs in driver error messages, e.g.
+       // "connection to postgresql://superset:hunter2@db:5432/superset 
failed".
+       // The user component may be empty (redis://:password@host). The 
password
+       // match is greedy up to the last @ in the whitespace-delimited token, 
so a
+       // malformed password containing a raw @ is masked in full rather than
+       // leaking its tail after the first @.
+       uriUserinfoRe = 
regexp.MustCompile(`([a-zA-Z][a-zA-Z0-9+.-]*://[^/@\s:]*):([^\s]+)@`)
+
+       // authorizationHeaderRe matches HTTP Authorization header values, with 
or
+       // without an auth scheme: "Authorization: Bearer eyJ...", 
"authorization=
+       // Basic dXNlcg==". The scheme word is kept; only the credential is 
masked.
+       authorizationHeaderRe = 
regexp.MustCompile(`(?i)\b(authorization\s*[=:]\s*(?:(?:bearer|basic|token|digest|negotiate)\s+)?)(\S+)`)
+
+       // bareBearerRe matches a bearer token without an Authorization prefix
+       // ("bearer eyJhbGci..."). The token must contain at least one digit so
+       // prose like "bearer authentication failed" is not mangled; real tokens
+       // (JWTs, base64, hex) virtually always contain digits.
+       bareBearerRe = 
regexp.MustCompile(`(?i)\b(bearer\s+)([A-Za-z0-9._~+/=-]*[0-9][A-Za-z0-9._~+/=-]*)`)
+
+       // credentialAssignmentRe matches free-form credential assignments such 
as
+       // "password=hunter2", "PGPASSWORD: hunter2", or "secret_key = hunter2".
+       // The key may carry prefixes/suffixes (DB_PASSWORD, api-key-prod).
+       credentialAssignmentRe = 
regexp.MustCompile(`(?i)([A-Za-z0-9_-]*(?:password|passwd|pwd|secret|token|api[_-]?key|credential|passphrase)[A-Za-z0-9_-]*\s*[=:]\s*)(\S+)`)

Review Comment:
   Before hand-rolling regex redaction, could you evaluate the existing Go 
libraries in this space and note the outcome in the PR description? The two 
closest that I found are [`masq`](https://github.com/m-mizutani/masq) and 
[`logfusc`](https://github.com/AngusGMorrison/logfusc).
   
   My read is that neither quite fits, but it's worth stating rather than 
implying:
   
   - `logfusc` wraps secrets the program holds as typed values (`Secret[T]`) so 
they redact on log/marshal. Here the credentials arrive inline in opaque 
task-container stdout — we never hold them as typed values, so there's nothing 
to wrap.
   - `masq` is a `slog` `ReplaceAttr` filter. This redaction targets CRD 
`status` and Kubernetes Events, not a log sink, and its pattern matching would 
still need the same regexes supplied via `WithRegex` — so it adds a dependency 
and a slog impedance mismatch without removing the regexes.
   
   If that's the conclusion, a sentence or two in the PR description 
("evaluated masq/logfusc; both target secrets held as typed values or slog 
attributes, whereas we scrub unknown credentials out of opaque subprocess 
output written to status/Events") would preempt exactly this question from 
other reviewers. If one of them *does* cover it, even better — we avoid 
maintaining regexes.
   
   Worth also noting the patterns here are inspired-by / reducible-to 
[gitleaks](https://github.com/gitleaks/gitleaks)' ruleset if you want a 
reference catalog, though importing gitleaks itself would be far too heavy for 
scrubbing a 256-char string.



-- 
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.

To unsubscribe, e-mail: [email protected]

For queries about this service, please contact Infrastructure at:
[email protected]


---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]

Reply via email to