feat(infra): convert sh-openswe-traces to HCP Terraform (#8)

Freeze SAM CD and add terraform/ for seahaven-prod with account-suffixed
buckets so HCP owns deploy before mgmt cutover. Suppress CKV_AWS_40 for
the LangSmith export IAM user with documented mitigations.
This commit is contained in:
Adam Moussa 2026-08-05 18:33:03 -04:00 • committed by GitHub
parent 50ac3de884
commit 5f430d9dfd
No known key found for this signature in database
GPG key ID: B5690EEEBB952194
16 changed files with 533 additions and 53 deletions

View file

@ -1,7 +1,14 @@
# FROZEN for PLAT-73: HCP Terraform is the sole deploy path.
# Previous SAM CD preserved as deploy.yaml.frozen until mgmt cutover.
# Do not re-enable until the migration is rolled back intentionally.
name: Deploy
on:
push:
branches: [main]
workflow_dispatch:
inputs:
confirm_unfreeze:
description: "Type UNFREEZE to acknowledge this bypasses PLAT-73 freeze"
required: true
type: string
permissions:
id-token: write
@ -13,6 +20,7 @@ concurrency:
jobs:
deploy:
if: ${{ inputs.confirm_unfreeze == 'UNFREEZE' }}
uses: Sea-Haven-Industries/.github/.github/workflows/cd-sam.yaml@3f746774229d41770727e2e4fd63ed5f5555a8b3 # v1.0.3
with:
stack-name: sh-openswe-traces

22
.github/workflows/deploy.yaml.frozen vendored Normal file
View file

@ -0,0 +1,22 @@
name: Deploy
on:
push:
branches: [main]
permissions:
id-token: write
contents: read
concurrency:
group: deploy
cancel-in-progress: false
jobs:
deploy:
uses: Sea-Haven-Industries/.github/.github/workflows/cd-sam.yaml@3f746774229d41770727e2e4fd63ed5f5555a8b3 # v1.0.3
with:
stack-name: sh-openswe-traces
cfn-role-arn: arn:aws:iam::328440206208:role/sh-openswe-traces-cfn-exec-role
secrets:
deploy-role-arn: ${{ secrets.AWS_DEPLOY_ROLE_ARN }}
parameter-overrides: ${{ secrets.SAM_PARAMETER_OVERRIDES }}

11
.gitignore vendored
View file

@ -5,3 +5,14 @@ samconfig.toml
*.pyc
__pycache__/
.DS_Store
# Local Terraform
terraform/.terraform/
*.tfstate
*.tfstate.*
crash.log
override.tf
override.tf.json
*_override.tf
*_override.tf.json
.terraformrc
terraform.rc

View file

@ -6,7 +6,11 @@
},
{
"id": "checkov-CKV_AWS_111-13",
"justification": "False positive. Same ExecRole. CKV_AWS_111 (write without constraint on Resource:*) fires on the KMS block (kms:CreateKey/CreateAlias), which is inherent to key creation — not-yet-existent keys cannot be ARN-scoped — with blast radius bounded to this stack's deploys. The other Resource:* statement is a Deny (flagging a Deny as over-permissive is nonsensical). CFN-only assumable, SourceAccount-conditioned. Verified proof-or-kill 2026-07-13."
"justification": "False positive. Same ExecRole. CKV_AWS_111 (write without constraint on Resource:*) fires on the KMS block (kms:CreateKey/CreateAlias), which is inherent to key creation \u2014 not-yet-existent keys cannot be ARN-scoped \u2014 with blast radius bounded to this stack's deploys. The other Resource:* statement is a Deny (flagging a Deny as over-permissive is nonsensical). CFN-only assumable, SourceAccount-conditioned. Verified proof-or-kill 2026-07-13."
},
{
"id": "checkov-CKV_AWS_40-10",
"justification": "Accepted risk for LangSmith Bulk Export (PLAT-73). CKV_AWS_40 prefers groups/roles over inline user policies; LangSmith's S3 destination requires long-lived IAM user access keys, so an IAM user with a single inline write-only policy is required. Mitigations: PutObject/AbortMultipartUpload only (no GetObject/DeleteObject); KMS Encrypt/GenerateDataKey/Decrypt gated by kms:ViaService=s3; access key minted out-of-band and never stored in Terraform state; secret on aws/secretsmanager key separate from the data CMK. Same pattern as the prior SAM LangSmithExportUser. REVISIT if LangSmith supports role assumption for Bulk Export destinations."
}
]
}

107
README.md
View file

@ -1,10 +1,13 @@
# sh-openswe-traces
LangSmith Bulk Export destination for the Open SWE deployment. A **storage-only**
SAM stack — no compute. LangSmith runs the export on its own schedule and writes
stack — no compute. LangSmith runs the export on its own schedule and writes
Parquet run/trace data into this bucket; we retain it for periodic auditing and
prompt/instruction improvement (query with Athena).
Owned by **HCP Terraform** in seahaven-prod (`sh-openswe-traces-prod`). Migrated
from mgmt SAM under [PLAT-73](https://seahaven.atlassian.net/browse/PLAT-73).
## Why this exists
LangSmith retains traces for ~14 days. To audit agent behavior over longer
@ -18,7 +21,8 @@ this repo is just the destination and its access controls.
LangSmith (Bulk Export, scheduled LangSmith-side)
│ s3:PutObject (IAM user access key, least-privilege)
▼
s3://sh-openswe-traces/langsmith/… SSE-KMS (alias/sh-openswe-traces)
s3://sh-openswe-traces-<account-id>/langsmith/…
│ SSE-KMS (alias/sh-openswe-traces)
│ lifecycle: → DEEP_ARCHIVE @ 90d
▼
Athena / manual audit
@ -26,64 +30,64 @@ LangSmith (Bulk Export, scheduled LangSmith-side)
| Resource | Name | Notes |
|---|---|---|
| S3 bucket | `sh-openswe-traces` | BPA all-on, SSE-KMS (default), versioned, TLS-only, `Retain` on delete |
| Log bucket | `sh-openswe-traces-logs` | S3 server access logs (SSE-S3) for read attribution; logs expire 365d |
| S3 bucket | `sh-openswe-traces-<account-id>` | BPA all-on, SSE-KMS (default), versioned, TLS-only, `prevent_destroy` |
| Log bucket | `sh-openswe-traces-logs-<account-id>` | S3 server access logs (SSE-S3); logs expire 365d |
| KMS CMK | `alias/sh-openswe-traces` | Rotation on; bucket default + writer encrypt through it |
| IAM user | auto-named (tag `sh-openswe-langsmith-export`) | Write-only LangSmith writer; `PutObject` bucket-wide, no read/delete |
| Secret | `sh-openswe/langsmith-export-s3` | Writer's access key (aws/secretsmanager key); populated post-deploy |
| IAM user | `sh-openswe-langsmith-export` | Write-only LangSmith writer; `PutObject` bucket-wide, no read/delete |
| Secret | `sh-openswe/langsmith-export-s3` | Writer's access key (aws/secretsmanager key); populated post-apply |
The IAM user is **not** given an explicit name so the app stack deploys under
`CAPABILITY_IAM`. It is referenced by ARN.
Account-suffixed bucket names avoid global S3 name collision with the retired mgmt
buckets during cutover.
## Deploy
## Deploy (HCP Terraform)
Two stacks:
- **`bootstrap.yaml`** → `sh-openswe-traces-bootstrap` — the CI IAM roles (OIDC deploy
role + a least-privilege CFN exec role). Deployed **once, manually, under admin**
(`CAPABILITY_NAMED_IAM`); rarely changes. Kept separate so CI never touches the shared
`github-cfn-execution-role`, which is roles-only and can't create this stack's KMS key,
secret, or IAM user.
- **`template.yaml`** → `sh-openswe-traces` — the app (buckets, CMK, writer, secret). CI
(`cd-sam.yaml`) deploys it on merge to `main`, assuming the OIDC deploy role and passing
`sh-openswe-traces-cfn-exec-role` as `--role-arn`.
Workspace: `sh-openswe-traces-prod` (org `seahaven`, project `seahaven-prod`).
Working directory: `terraform/`. Apply method: **Manual** until sealed.
Roles: `hcptf-sh-openswe-traces-plan` / `hcptf-sh-openswe-traces` (org-baseline
terraform-substrate).
```bash
# One-time bootstrap (admin):
aws cloudformation deploy --template-file bootstrap.yaml \
--stack-name sh-openswe-traces-bootstrap --capabilities CAPABILITY_NAMED_IAM \
--region us-east-1
# App stack — CI does this on merge to main; for a local/admin deploy:
cp samconfig.toml.example samconfig.toml
sam validate --lint && sam build
sam deploy --capabilities CAPABILITY_IAM --resolve-s3
cd terraform
terraform init
terraform plan
# First apply is Manual from the HCP UI (or terraform apply after plan confirm).
```
### Wire CI (after bootstrap + repo exist)
If the secret shell was created out-of-band (checklist step 1), import before the
first apply that manages it:
- Set repo secret **`AWS_DEPLOY_ROLE_ARN`** to the bootstrap `DeployRoleArn` output
(`arn:aws:iam::328440206208:role/githubdeploy-sh-openswe-traces`).
- `deploy.yaml` already passes `cfn-role-arn = sh-openswe-traces-cfn-exec-role`; the OIDC
trust is pinned to this repo's `main` ref.
```bash
terraform import aws_secretsmanager_secret.export_key \
'arn:aws:secretsmanager:us-east-1:011934824531:secret:sh-openswe/langsmith-export-s3-OQoqqU'
```
### Post-deploy: mint and store the writer access key
### Workspace env (once)
Workspace-level only (never project-scoped variable sets):
- `TFC_AWS_PROVIDER_AUTH=true`
- `TFC_AWS_PLAN_ROLE_ARN=arn:aws:iam::011934824531:role/hcptf-sh-openswe-traces-plan`
- `TFC_AWS_APPLY_ROLE_ARN=arn:aws:iam::011934824531:role/hcptf-sh-openswe-traces`
### Post-apply: mint and store the writer access key
The stack creates the IAM user and an empty secret; the access key is minted
out-of-band so it never lands in CloudFormation state. **Run this only on a trusted
out-of-band so it never lands in Terraform state. **Run this only on a trusted
single-user workstation, never in CI** — it handles a live credential.
```bash
USER=$(aws cloudformation describe-stacks --stack-name sh-openswe-traces \
--query "Stacks[0].Outputs[?OutputKey=='ExportUserName'].OutputValue" --output text)
USER=$(terraform -chdir=terraform output -raw export_user_name)
KEY_JSON=$(aws iam create-access-key --user-name "$USER" \
KEY_JSON=$(aws iam create-access-key --user-name "$USER" --profile seahaven-prod \
| jq '{AccessKeyId: .AccessKey.AccessKeyId, SecretAccessKey: .AccessKey.SecretAccessKey}')
# Pass the secret via stdin, not argv, so it never appears in the process table.
aws secretsmanager put-secret-value \
# Strip trailing newlines — a trailing \n breaks HTTP headers at runtime.
printf '%s' "$KEY_JSON" | aws secretsmanager put-secret-value \
--secret-id sh-openswe/langsmith-export-s3 \
--secret-string file:///dev/stdin <<<"$KEY_JSON"
--secret-string file:///dev/stdin \
--profile seahaven-prod \
--region us-east-1
unset KEY_JSON
```
@ -97,7 +101,9 @@ destination call validates by test-writing to the bucket.
**1. Create the destination** (creds pulled from Secrets Manager, never pasted):
```bash
BUCKET=$(terraform -chdir=terraform output -raw bucket_name)
CREDS=$(aws secretsmanager get-secret-value --secret-id sh-openswe/langsmith-export-s3 \
--profile seahaven-prod --region us-east-1 \
--query SecretString --output text)
AKID=$(jq -r .AccessKeyId <<<"$CREDS"); SAK=$(jq -r .SecretAccessKey <<<"$CREDS")
@ -105,7 +111,7 @@ curl -sS -X POST 'https://api.smith.langchain.com/api/v1/bulk-exports/destinatio
-H 'Content-Type: application/json' -H "X-API-Key: $LS_API_KEY" -H "X-Tenant-Id: $LS_TENANT" \
--data @- <<JSON | jq .
{ "destination_type": "s3", "display_name": "sh-openswe-traces us-east-1",
"config": { "bucket_name": "sh-openswe-traces", "prefix": "langsmith", "region": "us-east-1" },
"config": { "bucket_name": "$BUCKET", "prefix": "langsmith", "region": "us-east-1" },
"credentials": { "access_key_id": "$AKID", "secret_access_key": "$SAK" } }
JSON
```
@ -162,10 +168,10 @@ curl -sS 'https://api.smith.langchain.com/api/v1/bulk-exports' \
secret + LangSmith destination (`PATCH`/recreate), then delete the old key. Key age is
monitored account-wide by the Security Hub `ACCESS_KEYS_ROTATED` Config rule (flags at
90d) — rotation keeps it compliant.
- **Audit**: point Athena at `s3://sh-openswe-traces/langsmith/` (Parquet). Objects older
than 90 days are in Deep Archive — restore before querying.
- **Read attribution**: object access is logged to `s3://sh-openswe-traces-logs/s3-access/`
(S3 server access logging).
- **Audit**: point Athena at `s3://sh-openswe-traces-<account-id>/langsmith/` (Parquet).
Objects older than 90 days are in Deep Archive — restore before querying.
- **Read attribution**: object access is logged to
`s3://sh-openswe-traces-logs-<account-id>/s3-access/` (S3 server access logging).
- **Cost**: Deep Archive ≈ $1/TB/mo; expect the archive to dominate storage cost.
## Gotchas (from IAM cross-review)
@ -186,9 +192,10 @@ Traces can contain source code and secrets surfaced in tool I/O. Controls: SSE-K
rest (customer-managed CMK, bucket default), versioning (overwrite recovery), Block Public
Access, TLS-only bucket policy, a write-only least-privilege writer (bucket-wide `PutObject`
+ `kms:Decrypt` gated to `kms:ViaService=s3`, no read/delete), the credential secret on a
separate managed key, and S3 access logging. CI deploys via a dedicated least-privilege
exec role (see `bootstrap.yaml`), not the shared execution role.
separate managed key, and S3 access logging.
The IAM surface passed GPT-4.1 cross-review and a `/sh-security-review` fan-out +
proof-or-kill verifier (no blocking findings). Deferred, non-blocking: CloudTrail S3
data-events (org trail carries none; access logging covers attribution for now).
SAM `template.yaml` / `bootstrap.yaml` and the frozen GitHub `deploy.yaml` are retained
only until mgmt cutover completes; do not re-enable SAM CD.
Deferred, non-blocking: CloudTrail S3 data-events (org trail carries none; access logging
covers attribution for now).

29
terraform/.terraform.lock.hcl generated Normal file
View file

@ -0,0 +1,29 @@
# This file is maintained automatically by "terraform init".
# Manual edits may be lost in future updates.
provider "registry.terraform.io/hashicorp/aws" {
version = "6.58.0"
constraints = "~> 6.57"
hashes = [
"h1:1im6ypdeXLXq3sHElserTE62qmIqNiHLAyC3R5EnFFw=",
"h1:2kpake4zZKRX5437QVIRU3qFYH6Bjw/QE1fgVCPOrUg=",
"h1:OWl47Bo8Vzlf5srTUCmA6v4kvQGfah/P1joRtIYUUMc=",
"h1:UFot9S97tuAPvjKvoxm08sDG/gKYdDK+lMwsZKtLieY=",
"zh:1221253beee5629fb503d79cebc9bc661279cbc4be5d01db9ab4c1b702108250",
"zh:132bd0925bdc4b72446ac750b7ccb1e19b9ba8fbb6df57b2c1423314d2195d4f",
"zh:18cda250b9e82b753808715893c8927f132273c00ffae7a697d65ac1cb577e48",
"zh:204c944f1fb7f440a335bb2083c9691a9d1f677aea9701025dd5816aee41f0ba",
"zh:2dc41df289f2b10a01e650cdd73699955f0ab0645d09cfb114a8cd0f4cc4ede7",
"zh:345633dfa9a234659d52aadd126e6dce658518c3ab5cbf6d871221287ed5ec56",
"zh:4dadcced73e742903158bc9838936d911f3fa4c2c37b5591c1a28f8f2a1902a6",
"zh:5bc60cc2b8c093da98b211d9f6c21c9ecec0f21b944d9ecbe3961fba33086e80",
"zh:6cc8f084938b0033a9c0c910989919dad6b1683e76e0afa1a5c604e39f398a75",
"zh:7db214647f79de9a033b5dfd6cbfaa42d53c4d056b32cb3acc7ad99306dd548a",
"zh:9078589ec881cee7ed9403af262c98ff257fb3e1baae72ff6429a398b1c730af",
"zh:9b12af85486a96aedd8d7984b0ff811a4b42e3d88dad1a3fb4c0b580d04fa425",
"zh:bd5bce6aec4d4922b1127b8575688bd4bc4279670ee28d198ede404709826c7c",
"zh:cd900ecf56d21023873898b06e40234f3f4d350f2343b7d9b980d6c5cb604fae",
"zh:dbe93b276a84421026b956c3c5b4eb8897da6cbb54b93cb89f5d6ebbd30805ca",
"zh:f6b6c7bb2dbf04ee085e5c22f7a65b3ccaebf368ed95584dc0dfee8a22771056",
]
}

48
terraform/iam.tf Normal file
View file

@ -0,0 +1,48 @@
resource "aws_iam_user" "langsmith_export" {
name = local.export_user_name
tags = {
Name = "sh-openswe-langsmith-export"
purpose = "langsmith-bulk-export-writer"
}
}
resource "aws_iam_user_policy" "langsmith_export_put" {
name = "langsmith-export-put"
user = aws_iam_user.langsmith_export.name
# Bucket-wide PutObject (not prefix-scoped): LangSmith destination validation
# writes a test object whose key is not guaranteed under export_prefix.
# Deliberately NO s3:GetObject / s3:DeleteObject — write-only.
policy = jsonencode({
Version = "2012-10-17"
Statement = [
{
Sid = "PutExportObjects"
Effect = "Allow"
Action = [
"s3:PutObject",
"s3:AbortMultipartUpload",
]
Resource = "${aws_s3_bucket.traces.arn}/*"
},
{
Sid = "EncryptWithBucketKey"
Effect = "Allow"
Action = [
"kms:GenerateDataKey",
"kms:Encrypt",
# Required by S3 at CompleteMultipartUpload for SSE-KMS. Safe: the
# writer has no s3:GetObject, so there is no object to decrypt/exfil.
"kms:Decrypt",
]
Resource = aws_kms_key.traces.arn
Condition = {
StringEquals = {
"kms:ViaService" = "s3.${var.aws_region}.amazonaws.com"
}
}
}
]
})
}

25
terraform/kms.tf Normal file
View file

@ -0,0 +1,25 @@
resource "aws_kms_key" "traces" {
description = "SSE-KMS CMK for sh-openswe-traces (LangSmith export archive)."
enable_key_rotation = true
deletion_window_in_days = 30
policy = jsonencode({
Version = "2012-10-17"
Statement = [
{
Sid = "EnableIAMPolicies"
Effect = "Allow"
Principal = {
AWS = "arn:aws:iam::${local.account_id}:root"
}
Action = "kms:*"
Resource = "*"
}
]
})
}
resource "aws_kms_alias" "traces" {
name = local.kms_alias_name
target_key_id = aws_kms_key.traces.key_id
}

14
terraform/locals.tf Normal file
View file

@ -0,0 +1,14 @@
locals {
account_id = data.aws_caller_identity.current.account_id
# Account-suffixed names so seahaven-prod can apply in parallel with mgmt
# teardown (S3 bucket names are globally unique; PLAT-73 locked decision).
traces_bucket_name = "sh-openswe-traces-${local.account_id}"
logs_bucket_name = "sh-openswe-traces-logs-${local.account_id}"
export_user_name = "sh-openswe-langsmith-export"
secret_name = "sh-openswe/langsmith-export-s3"
kms_alias_name = "alias/sh-openswe-traces"
}
data "aws_caller_identity" "current" {}

36
terraform/outputs.tf Normal file
View file

@ -0,0 +1,36 @@
output "bucket_name" {
value = aws_s3_bucket.traces.id
description = "LangSmith Bulk Export destination bucket (account-suffixed)"
}
output "bucket_arn" {
value = aws_s3_bucket.traces.arn
}
output "logs_bucket_name" {
value = aws_s3_bucket.logs.id
}
output "kms_key_arn" {
value = aws_kms_key.traces.arn
}
output "kms_alias" {
value = aws_kms_alias.traces.name
}
output "export_user_name" {
value = aws_iam_user.langsmith_export.name
}
output "export_user_arn" {
value = aws_iam_user.langsmith_export.arn
}
output "export_key_secret_name" {
value = aws_secretsmanager_secret.export_key.name
}
output "export_key_secret_arn" {
value = aws_secretsmanager_secret.export_key.arn
}

11
terraform/providers.tf Normal file
View file

@ -0,0 +1,11 @@
provider "aws" {
region = var.aws_region
default_tags {
tags = {
Project = "sh-openswe-traces"
ManagedBy = "terraform"
Workspace = "sh-openswe-traces-prod"
}
}
}

219
terraform/s3.tf Normal file
View file

@ -0,0 +1,219 @@
resource "aws_s3_bucket" "logs" {
bucket = local.logs_bucket_name
tags = {
role = "s3-access-logs"
}
lifecycle {
prevent_destroy = true
}
}
resource "aws_s3_bucket_public_access_block" "logs" {
bucket = aws_s3_bucket.logs.id
block_public_acls = true
block_public_policy = true
ignore_public_acls = true
restrict_public_buckets = true
}
resource "aws_s3_bucket_ownership_controls" "logs" {
bucket = aws_s3_bucket.logs.id
rule {
object_ownership = "BucketOwnerEnforced"
}
}
resource "aws_s3_bucket_server_side_encryption_configuration" "logs" {
bucket = aws_s3_bucket.logs.id
rule {
apply_server_side_encryption_by_default {
# SSE-S3 only: S3 log delivery cannot write to an SSE-KMS bucket.
sse_algorithm = "AES256"
}
}
}
resource "aws_s3_bucket_lifecycle_configuration" "logs" {
bucket = aws_s3_bucket.logs.id
rule {
id = "expire-access-logs"
status = "Enabled"
filter {}
expiration {
days = 365
}
}
}
resource "aws_s3_bucket_policy" "logs" {
bucket = aws_s3_bucket.logs.id
policy = jsonencode({
Version = "2012-10-17"
Statement = [
{
Sid = "S3ServerAccessLogsWrite"
Effect = "Allow"
Principal = {
Service = "logging.s3.amazonaws.com"
}
Action = "s3:PutObject"
Resource = "${aws_s3_bucket.logs.arn}/s3-access/*"
Condition = {
ArnLike = {
"aws:SourceArn" = aws_s3_bucket.traces.arn
}
StringEquals = {
"aws:SourceAccount" = local.account_id
}
}
},
{
Sid = "DenyInsecureTransport"
Effect = "Deny"
Principal = "*"
Action = "s3:*"
Resource = [
aws_s3_bucket.logs.arn,
"${aws_s3_bucket.logs.arn}/*",
]
Condition = {
Bool = {
"aws:SecureTransport" = "false"
}
}
}
]
})
}
resource "aws_s3_bucket" "traces" {
bucket = local.traces_bucket_name
lifecycle {
prevent_destroy = true
}
}
resource "aws_s3_bucket_public_access_block" "traces" {
bucket = aws_s3_bucket.traces.id
block_public_acls = true
block_public_policy = true
ignore_public_acls = true
restrict_public_buckets = true
}
resource "aws_s3_bucket_ownership_controls" "traces" {
bucket = aws_s3_bucket.traces.id
rule {
object_ownership = "BucketOwnerEnforced"
}
}
resource "aws_s3_bucket_versioning" "traces" {
bucket = aws_s3_bucket.traces.id
versioning_configuration {
status = "Enabled"
}
}
resource "aws_s3_bucket_server_side_encryption_configuration" "traces" {
bucket = aws_s3_bucket.traces.id
rule {
apply_server_side_encryption_by_default {
sse_algorithm = "aws:kms"
kms_master_key_id = aws_kms_key.traces.arn
}
bucket_key_enabled = true
}
}
resource "aws_s3_bucket_lifecycle_configuration" "traces" {
bucket = aws_s3_bucket.traces.id
rule {
id = "archive-exports-to-deep-archive"
status = "Enabled"
filter {
prefix = var.export_prefix
}
transition {
days = 90
storage_class = "DEEP_ARCHIVE"
}
}
rule {
id = "expire-noncurrent-versions"
status = "Enabled"
filter {}
noncurrent_version_expiration {
noncurrent_days = 90
}
}
rule {
id = "abort-incomplete-multipart"
status = "Enabled"
filter {}
abort_incomplete_multipart_upload {
days_after_initiation = 7
}
}
}
resource "aws_s3_bucket_logging" "traces" {
bucket = aws_s3_bucket.traces.id
target_bucket = aws_s3_bucket.logs.id
target_prefix = "s3-access/"
depends_on = [aws_s3_bucket_policy.logs]
}
resource "aws_s3_bucket_policy" "traces" {
bucket = aws_s3_bucket.traces.id
# No SSE-header enforcement Deny. LangSmith's exporter does not send an
# "aws:kms" SSE header, so a "must be aws:kms" Deny blocks its writes — and
# StringNotEqualsIfExists on a Deny also blocks header-less puts. Encryption
# is guaranteed by the bucket DEFAULT (SSE-KMS with our CMK).
policy = jsonencode({
Version = "2012-10-17"
Statement = [
{
Sid = "DenyInsecureTransport"
Effect = "Deny"
Principal = "*"
Action = "s3:*"
Resource = [
aws_s3_bucket.traces.arn,
"${aws_s3_bucket.traces.arn}/*",
]
Condition = {
Bool = {
"aws:SecureTransport" = "false"
}
}
}
]
})
}

15
terraform/secrets.tf Normal file
View file

@ -0,0 +1,15 @@
# Holder only — the access key is minted out-of-band and written with
# `aws secretsmanager put-secret-value` (see README). Never manage secret
# values in Terraform so they never enter HCP state.
#
# Pre-created in seahaven-prod for PLAT-73 step 1. Import on first apply:
# terraform import aws_secretsmanager_secret.export_key \
# arn:aws:secretsmanager:us-east-1:011934824531:secret:sh-openswe/langsmith-export-s3-OQoqqU
resource "aws_secretsmanager_secret" "export_key" {
name = local.secret_name
description = "Access key for the sh-openswe-langsmith-export IAM user, consumed by LangSmith Bulk Export. Populated out-of-band post-deploy; rotate quarterly."
# Encrypted with the aws/secretsmanager managed key — deliberately NOT the
# trace CMK, so the credential and the data it protects never share a key
# the writer principal holds any KMS grant on.
}

View file

@ -0,0 +1,2 @@
aws_region = "us-east-1"
export_prefix = "langsmith/"

11
terraform/variables.tf Normal file
View file

@ -0,0 +1,11 @@
variable "aws_region" {
type = string
description = "AWS region for all resources"
default = "us-east-1"
}
variable "export_prefix" {
type = string
description = "S3 key prefix LangSmith writes exports under (also the lifecycle scope)"
default = "langsmith/"
}

18
terraform/versions.tf Normal file
View file

@ -0,0 +1,18 @@
terraform {
required_version = ">= 1.7.0"
required_providers {
aws = {
source = "hashicorp/aws"
version = "~> 6.57"
}
}
cloud {
organization = "seahaven"
workspaces {
name = "sh-openswe-traces-prod"
}
}
}