一键在 Manus 中运行任何 Skill

gcp-logs

星标66

分支597

更新时间2026年1月28日 10:43

Read and query GCP logs for Hyperlane agents using gcloud CLI. Use when investigating relayer, validator, or scraper logs, debugging message processing, or analyzing operational issues. Provides efficient filtering and context management strategies.

安装

用 Codex 或 Claude 帮你安装复制这段 Prompt，粘贴到 Codex、Claude 或其他助手里，让它检查 Skill 页面并帮你完成安装。

在 Manus 中运行

来源

hyperlane-xyz

hyperlane-xyz/hyperlane-monorepo

打开 GitHub 仓库查看创作者相关仓库

下载

在 Manus 中运行

GCP Logs Query Skill

When to Use

Investigating relayer, validator, or scraper behavior
Debugging message processing issues
Analyzing operational incidents
Looking for specific errors or patterns in agent logs

Prerequisites

gcloud CLI installed and authenticated
Access to abacus-labs-dev GCP project

Base Query Templates

Relayer (Omniscient)

gcloud logging read 'resource.type="k8s_container" AND resource.labels.project_id="abacus-labs-dev" AND resource.labels.location="us-east1-c" AND resource.labels.cluster_name="hyperlane-mainnet" AND resource.labels.namespace_name="mainnet3" AND labels.k8s-pod/app_kubernetes_io/component="relayer" AND labels.k8s-pod/app_kubernetes_io/instance="omniscient-relayer" AND labels.k8s-pod/app_kubernetes_io/name="hyperlane-agent"' --project=abacus-labs-dev --limit=50 --format=json --freshness=1d

Validator

gcloud logging read 'resource.type="k8s_container" AND resource.labels.project_id="abacus-labs-dev" AND resource.labels.location="us-east1-c" AND resource.labels.cluster_name="hyperlane-mainnet" AND resource.labels.namespace_name="mainnet3" AND labels.k8s-pod/app_kubernetes_io/component="validator" AND labels.k8s-pod/app_kubernetes_io/name="hyperlane-agent"' --project=abacus-labs-dev --limit=50 --format=json --freshness=1d

Scraper

gcloud logging read 'resource.type="k8s_container" AND resource.labels.project_id="abacus-labs-dev" AND resource.labels.location="us-east1-c" AND resource.labels.cluster_name="hyperlane-mainnet" AND resource.labels.namespace_name="mainnet3" AND labels.k8s-pod/app_kubernetes_io/component="scraper3" AND labels.k8s-pod/app_kubernetes_io/instance="omniscient-scraper" AND labels.k8s-pod/app_kubernetes_io/name="hyperlane-agent"' --project=abacus-labs-dev --limit=50 --format=json --freshness=1d

Noise Filtering

Add these filters to reduce noisy log lines that consume context without providing value:

-jsonPayload.fields.message="Found log(s) in index range"
-jsonPayload.fields.message="Dispatching get_public_key"
NOT "Instantiated AWS signer"
-jsonPayload.fields.message="Ingesting leaf"
-jsonPayload.fields.message="Message already marked as processed in DB"
-jsonPayload.fields.message="Message destined for self, skipping"
-jsonPayload.fields.message="Message has already been delivered, marking as submitted."
-jsonPayload.fields.message="Popped OpQueue operations"
-jsonPayload.fields.message="Validator returned latest index"
-jsonPayload.fields.message="Found signed checkpoint"
-jsonPayload.fields.return="Ok(None)"
-jsonPayload.fields.message="Fast forwarded current sequence"
-jsonPayload.fields.message="Cursor can't make progress, sleeping"
-jsonPayload.fields.message="fallback_request"
-jsonPayload.fields.message="No message found in DB for leaf index"
-jsonPayload.fields.message="Processor working on message"
-jsonPayload.fields.message="Message destined for unknown domain, skipping"

Progressive Query Strategy (Token Efficiency)

Step 1: Fetch Message Field Only First

To minimize context consumption, first fetch only the message field:

gcloud logging read '[BASE_QUERY] AND "[search_term]"' --project=abacus-labs-dev --limit=30 --format='json(jsonPayload.fields.message,timestamp)' --freshness=1d

This gives you a quick overview without the full log payload.

Step 2: Get Full Context for Specific Entries

Once you identify interesting log entries, fetch full details:

gcloud logging read '[BASE_QUERY] AND "[specific_identifier]"' --project=abacus-labs-dev --limit=20 --format=json --freshness=1d

Step 3: Extract Specific Fields

When you need specific details, use jq or grep to extract:

gcloud logging read '[QUERY]' --format=json | jq '.[].jsonPayload.fields.error'

Common Query Patterns

Search by Message ID

gcloud logging read '[BASE_QUERY] AND "0x[MESSAGE_ID]"' --project=abacus-labs-dev --limit=50 --format=json --freshness=1d

Search for Errors/Warnings

gcloud logging read '[BASE_QUERY] AND severity>="WARNING"' --project=abacus-labs-dev --limit=50 --format=json --freshness=1d

Search by Chain/Domain

gcloud logging read '[BASE_QUERY] AND jsonPayload.spans.domain:"[chain_name]"' --project=abacus-labs-dev --limit=50 --format=json --freshness=1d

Search for Stuck Messages (High Retry Count)

gcloud logging read '[BASE_QUERY] AND jsonPayload.fields.num_retries>=5' --project=abacus-labs-dev --limit=30 --format=json --freshness=1d

Search for Gas Estimation Errors

gcloud logging read '[BASE_QUERY] AND "eth_estimateGas"' --project=abacus-labs-dev --limit=30 --format=json --freshness=1d

Search by App Context

gcloud logging read '[BASE_QUERY] AND jsonPayload.fields.app_context:"[APP_CONTEXT]"' --project=abacus-labs-dev --limit=30 --format=json --freshness=1d

Time Range Options

--freshness=1h - Last hour
--freshness=1d - Last day
--freshness=7d - Last week
Or use explicit timestamps in filter: timestamp>="2026-01-27T00:00:00Z"

Output Format Options

--format=json - Full JSON (verbose, high context)
--format='json(jsonPayload.fields.message,timestamp)' - Specific fields only (efficient)
--format='value(jsonPayload.fields.message)' - Just values, no structure

Key Log Fields to Focus On

Field	Description
`jsonPayload.fields.message`	Main log message
`jsonPayload.fields.error`	Error details
`jsonPayload.spans[].domain`	Chain involved
`jsonPayload.fields.num_retries`	Retry count
`jsonPayload.fields.operations`	Pending message details
`jsonPayload.span.id`	Message ID in span context

Environment Variations

Environment	Namespace	Cluster
mainnet3	`mainnet3`	`hyperlane-mainnet`
testnet4	`testnet4`	`hyperlane-mainnet`

Tips

Always start specific - Search for exact message IDs or error patterns first
Use noise filters - The base logs are very noisy; always filter
Limit results - Use --limit to avoid overwhelming context
Progressive detail - Start with message field only, expand as needed
Time bound queries - Use --freshness or timestamp filters
Pipe to grep/jq - Post-process large results locally

同仓库更多 Skills

同仓库

set-moonpay-standing-quotes

hyperlane-xyz/hyperlane-monorepo

Update standing fee quotes for the CROSS/moonpay warp route. Use when asked to set, update, or change fees/bps for specific chains, tokens, or directions in the MoonPay route. Also use when asked to show or read current standing quotes.

2026-07-0366

hyperlane-message-utils

hyperlane-xyz/hyperlane-monorepo

Encode and decode raw Hyperlane messages to/from packed hex bytes using the CLI. Use when you need to construct a message for testing, inspect a raw message from logs or a transaction, or decode a warp transfer body.

2026-06-2966

hyperlane-utils

hyperlane-xyz/hyperlane-monorepo

Helper skill for using Hyperlane CLI address conversion utilities (addressToBytes32 and bytes32ToAddress) for cross-chain message formatting

2026-06-2966

warp-route-check

hyperlane-xyz/hyperlane-monorepo

A skill that allows a user to validate that a warp route deployment will be in the expected state after applying some configuration

2026-06-2966

warp-simulate-safe-txs

hyperlane-xyz/hyperlane-monorepo

Simulate pending Safe (multisig) governance txs for a warp route by replaying the literal calldata onto anvil forks, self-relaying any ICA messages, then running warp check against the desired registry config. Use to verify that a not-yet-signed Safe batch produces the intended warp route config before signing.

2026-06-0966

warp-deploy-extend-route

hyperlane-xyz/hyperlane-monorepo

Add a new chain to an existing warp route owned by a customer. Reads a Linear ticket, adds the new chain to the deploy.yaml, builds a customer-specific strategy file for existing chains, runs warp apply, and outputs transaction files for the customer to sign via their multisig.

2026-06-0866

name	gcp-logs
description	Read and query GCP logs for Hyperlane agents using gcloud CLI. Use when investigating relayer, validator, or scraper logs, debugging message processing, or analyzing operational issues. Provides efficient filtering and context management strategies.

GCP Logs Query Skill

When to Use

Investigating relayer, validator, or scraper behavior
Debugging message processing issues
Analyzing operational incidents
Looking for specific errors or patterns in agent logs

Prerequisites

gcloud CLI installed and authenticated
Access to abacus-labs-dev GCP project

Base Query Templates

Relayer (Omniscient)

gcloud logging read 'resource.type="k8s_container" AND resource.labels.project_id="abacus-labs-dev" AND resource.labels.location="us-east1-c" AND resource.labels.cluster_name="hyperlane-mainnet" AND resource.labels.namespace_name="mainnet3" AND labels.k8s-pod/app_kubernetes_io/component="relayer" AND labels.k8s-pod/app_kubernetes_io/instance="omniscient-relayer" AND labels.k8s-pod/app_kubernetes_io/name="hyperlane-agent"' --project=abacus-labs-dev --limit=50 --format=json --freshness=1d

Validator

gcloud logging read 'resource.type="k8s_container" AND resource.labels.project_id="abacus-labs-dev" AND resource.labels.location="us-east1-c" AND resource.labels.cluster_name="hyperlane-mainnet" AND resource.labels.namespace_name="mainnet3" AND labels.k8s-pod/app_kubernetes_io/component="validator" AND labels.k8s-pod/app_kubernetes_io/name="hyperlane-agent"' --project=abacus-labs-dev --limit=50 --format=json --freshness=1d

Scraper

gcloud logging read 'resource.type="k8s_container" AND resource.labels.project_id="abacus-labs-dev" AND resource.labels.location="us-east1-c" AND resource.labels.cluster_name="hyperlane-mainnet" AND resource.labels.namespace_name="mainnet3" AND labels.k8s-pod/app_kubernetes_io/component="scraper3" AND labels.k8s-pod/app_kubernetes_io/instance="omniscient-scraper" AND labels.k8s-pod/app_kubernetes_io/name="hyperlane-agent"' --project=abacus-labs-dev --limit=50 --format=json --freshness=1d

Noise Filtering

Add these filters to reduce noisy log lines that consume context without providing value:

-jsonPayload.fields.message="Found log(s) in index range"
-jsonPayload.fields.message="Dispatching get_public_key"
NOT "Instantiated AWS signer"
-jsonPayload.fields.message="Ingesting leaf"
-jsonPayload.fields.message="Message already marked as processed in DB"
-jsonPayload.fields.message="Message destined for self, skipping"
-jsonPayload.fields.message="Message has already been delivered, marking as submitted."
-jsonPayload.fields.message="Popped OpQueue operations"
-jsonPayload.fields.message="Validator returned latest index"
-jsonPayload.fields.message="Found signed checkpoint"
-jsonPayload.fields.return="Ok(None)"
-jsonPayload.fields.message="Fast forwarded current sequence"
-jsonPayload.fields.message="Cursor can't make progress, sleeping"
-jsonPayload.fields.message="fallback_request"
-jsonPayload.fields.message="No message found in DB for leaf index"
-jsonPayload.fields.message="Processor working on message"
-jsonPayload.fields.message="Message destined for unknown domain, skipping"

Progressive Query Strategy (Token Efficiency)

Step 1: Fetch Message Field Only First

To minimize context consumption, first fetch only the message field:

gcloud logging read '[BASE_QUERY] AND "[search_term]"' --project=abacus-labs-dev --limit=30 --format='json(jsonPayload.fields.message,timestamp)' --freshness=1d

This gives you a quick overview without the full log payload.

Step 2: Get Full Context for Specific Entries

Once you identify interesting log entries, fetch full details:

gcloud logging read '[BASE_QUERY] AND "[specific_identifier]"' --project=abacus-labs-dev --limit=20 --format=json --freshness=1d

Step 3: Extract Specific Fields

When you need specific details, use jq or grep to extract:

gcloud logging read '[QUERY]' --format=json | jq '.[].jsonPayload.fields.error'

Common Query Patterns

Search by Message ID

gcloud logging read '[BASE_QUERY] AND "0x[MESSAGE_ID]"' --project=abacus-labs-dev --limit=50 --format=json --freshness=1d

Search for Errors/Warnings

gcloud logging read '[BASE_QUERY] AND severity>="WARNING"' --project=abacus-labs-dev --limit=50 --format=json --freshness=1d

Search by Chain/Domain

gcloud logging read '[BASE_QUERY] AND jsonPayload.spans.domain:"[chain_name]"' --project=abacus-labs-dev --limit=50 --format=json --freshness=1d

Search for Stuck Messages (High Retry Count)

gcloud logging read '[BASE_QUERY] AND jsonPayload.fields.num_retries>=5' --project=abacus-labs-dev --limit=30 --format=json --freshness=1d

Search for Gas Estimation Errors

gcloud logging read '[BASE_QUERY] AND "eth_estimateGas"' --project=abacus-labs-dev --limit=30 --format=json --freshness=1d

Search by App Context

gcloud logging read '[BASE_QUERY] AND jsonPayload.fields.app_context:"[APP_CONTEXT]"' --project=abacus-labs-dev --limit=30 --format=json --freshness=1d

Time Range Options

--freshness=1h - Last hour
--freshness=1d - Last day
--freshness=7d - Last week
Or use explicit timestamps in filter: timestamp>="2026-01-27T00:00:00Z"

Output Format Options

--format=json - Full JSON (verbose, high context)
--format='json(jsonPayload.fields.message,timestamp)' - Specific fields only (efficient)
--format='value(jsonPayload.fields.message)' - Just values, no structure

Key Log Fields to Focus On

Field	Description
`jsonPayload.fields.message`	Main log message
`jsonPayload.fields.error`	Error details
`jsonPayload.spans[].domain`	Chain involved
`jsonPayload.fields.num_retries`	Retry count
`jsonPayload.fields.operations`	Pending message details
`jsonPayload.span.id`	Message ID in span context

Environment Variations

Environment	Namespace	Cluster
mainnet3	`mainnet3`	`hyperlane-mainnet`
testnet4	`testnet4`	`hyperlane-mainnet`

Tips

Always start specific - Search for exact message IDs or error patterns first
Use noise filters - The base logs are very noisy; always filter
Limit results - Use --limit to avoid overwhelming context
Progressive detail - Start with message field only, expand as needed
Time bound queries - Use --freshness or timestamp filters
Pipe to grep/jq - Post-process large results locally