Agent-Skills.md

Agent Skills: Manage Workloads

**Autonomously create** filters, throttles, and classification rules to implement new workloads or manage workload lifecycle in response to changing operational needs

UncategorizedID: teradata-labs/claude-cookbooks/manage-workloads

Repository

teradata-labs/claude-cookbooks
teradata-labsLicense: Apache-2.0
6

Install this agent skill to your local

pnpm dlx add-skill https://github.com/teradata-labs/claude-cookbooks/tree/HEAD/skills/operations/workloadmanagement/control/manage-workloads

Skill Files

Browse the full folder contents for manage-workloads.

Download Skill

Loading file tree…

skills/operations/workloadmanagement/control/manage-workloads/SKILL.md

Skill Metadata

Name
manage-workloads
Description
**Autonomously create** filters, throttles, and classification rules to implement new workloads or manage workload lifecycle in response to changing operational needs

Manage Workloads

Autonomously create and configure filters, throttles, and classification rules to implement new workloads, manage workload lifecycle, and dynamically control system behavior in response to operational changes.

🎯 New Autonomous Capabilities

This skill can now execute workload management, not just recommend it!

With tdwm-mcp v1.5.0, this skill has evolved from advisory to autonomous:

  • ✅ CREATE filters to route queries to workloads
  • ✅ CREATE throttles to protect workload resources
  • ✅ ENABLE/DISABLE filters and throttles for lifecycle management
  • ✅ USE TEMPLATES for common workload patterns
  • ✅ VALIDATE configurations before execution
  • ✅ VERIFY workload implementation

Note: Workload definitions themselves are created via Teradata TDWM utilities, but this skill automates the filter and throttle creation needed to make workloads functional.

Instructions

When to Use This Skill

  • Implementing new workload for application or user group
  • Creating filters to route queries to existing workload
  • Adding throttles to protect workload resources
  • Emergency situation requires temporary workload protection
  • Responding to maintenance windows or special events
  • Managing seasonal or special-purpose workload activation

Available MCP Tools

Monitoring & Analysis:

  • list_active_WD - List currently active workloads
  • list_WDs - List all workloads (active and inactive)
  • show_tdwm_summary - View workload distribution
  • show_tasm_statistics - Monitor workload effectiveness
  • list_rulesets - List all available rulesets

Configuration Management (NEW ✨):

  • create_filter_rule - Create filter to route queries to workload
  • create_system_throttle - Create throttle for workload protection
  • enable_filter_rule / disable_filter_rule - Lifecycle management
  • enable_throttle / disable_throttle - Lifecycle management
  • add_classification_to_rule - Add criteria to existing filter/throttle
  • activate_ruleset - Apply all pending changes

Available MCP Resources (NEW ✨)

Templates:

  • tdwm://templates/filter - List all filter templates
  • tdwm://template/filter/application-basic - Route by application
  • tdwm://template/filter/user-group - Route by user/account
  • tdwm://template/filter/time-based - Time-based routing
  • tdwm://templates/throttle - List all throttle templates
  • tdwm://template/throttle/application-basic - Basic workload throttle
  • tdwm://template/throttle/user-concurrency - Per-user limits

Reference Data:

  • tdwm://reference/classification-types - All 31 classification types
  • tdwm://reference/operators - Classification operators (I, O, IO)
  • tdwm://reference/filter-actions - Filter actions (ACCEPT/REJECT)
  • tdwm://reference/throttle-types - Throttle types (DM, M)

Discovery:

  • tdwm://rulesets - List all rulesets
  • tdwm://system/active-ruleset - Get currently active ruleset
  • tdwm://ruleset/{name}/filters - List filters in ruleset
  • tdwm://ruleset/{name}/throttles - List throttles in ruleset
  • tdwm://ruleset/{name}/pending-changes - Check pending changes

Workflows:

  • tdwm://workflow/create-filter - Step-by-step filter creation guide
  • tdwm://workflow/create-throttle - Step-by-step throttle creation guide

Step-by-Step Workflow

Phase 1: Discovery & Analysis (Read-Only)

  1. Understand Current State

    • Get active ruleset: tdwm://system/active-ruleset
    • List all workloads: list_active_WD and list_WDs
    • Review workload distribution: show_tdwm_summary
    • Document baseline before making changes

  2. Identify Workload Requirements

    • What queries should route to this workload?
    • How should queries be classified? (query band, user, account)
    • What concurrency limits are needed?
    • What priority level is appropriate?
    • Are there time-based or complexity requirements?

  3. Review Existing Configuration

    • List filters: tdwm://ruleset/{name}/filters
    • List throttles: tdwm://ruleset/{name}/throttles
    • Check for conflicts or overlaps with existing rules
    • Identify gaps in current configuration

  4. Plan Implementation

    • Which filters need to be created or enabled?
    • Which throttles need to be created or enabled?
    • What classification criteria should be used?
    • What is the activation sequence?

Phase 2: Execute Changes (Autonomous - NEW!)

  1. Select Approach: Template or Custom

    • Template-Based (Recommended):

      • Browse templates: tdwm://templates/filter and tdwm://templates/throttle
      • Select templates matching use case
      • Customize parameters for your workload

    • Custom Configuration:

      • Validate classification types: tdwm://reference/classification-types
      • Validate operators: tdwm://reference/operators
      • Design custom classification criteria

  2. Create Filter for Query Routing

    • Create new filter:

      Use: create_filter_rule(
  ruleset_name, filter_name,
  classification_criteria,
  action_type, workload_name
)

    • Filter routes queries to target workload based on criteria

  3. Create Throttle for Resource Protection

    • Create new throttle:

      Use: create_system_throttle(
  ruleset_name, throttle_name, throttle_type,
  limit, classification_criteria
)

    • Throttle limits concurrency for workload protection

  4. Enable and Activate

    • Enable filter: enable_filter_rule(ruleset_name, filter_name)
    • Enable throttle: enable_throttle(ruleset_name, throttle_name)
    • Check pending changes: tdwm://ruleset/{name}/pending-changes
    • Activate ruleset: activate_ruleset(ruleset_name)

Phase 3: Verification (Autonomous - NEW!)

  1. Verify Configuration

    • Read back filter: tdwm://ruleset/{name}/filter/{filter_name}
    • Read back throttle: tdwm://ruleset/{name}/throttle/{throttle_name}
    • Confirm enabled states are correct
    • Confirm classification criteria match expectations

  2. Monitor Impact

    • Wait 5-10 minutes for statistics to update
    • Check workload distribution: show_tdwm_summary
    • Verify queries routing to workload correctly
    • Monitor throttle statistics: show_trottle_statistics
    • Check TASM events: show_tasm_even_history

  3. Iterate if Needed

    • If routing not working: add more classification criteria
    • If throttle too restrictive: adjust limit
    • Use configuration tools for adjustments

Examples

Example 1: Implement New Application Workload (Autonomous)

Scenario: New BI tool launching Monday, need to route queries to BI_ANALYTICS workload

Discovery:

1. list_active_WD → BI_ANALYTICS workload exists (created by DBA)
2. show_query_band → BI tool sets 'APP=BI_DASHBOARD'
3. tdwm://system/active-ruleset → "Tactical"
4. tdwm://ruleset/Tactical/filters → No BI filter exists
5. show_tdwm_summary → BI queries currently in DEFAULT workload

Analysis:

  • Workload definition exists, needs filter for routing
  • BI tool sets query band (simple classification)
  • Need throttle to limit to 15 concurrent queries
  • Can use templates for quick implementation

Execution (Autonomous):

1. Get filter template:
   tdwm://template/filter/application-basic
   → Shows APPL-based filter pattern

2. Create filter to route BI queries:
   create_filter_rule(
     ruleset_name="Tactical",
     filter_name="BI_DASHBOARD_FILTER",
     classification_criteria=[{
       "description": "BI Dashboard Application",
       "type": "APPL",
       "value": "BI_DASHBOARD",
       "operator": "I"
     }],
     action_type="ACCEPT",
     workload_name="BI_ANALYTICS"
   )

3. Get throttle template:
   tdwm://template/throttle/application-basic
   → Shows APPL-based throttle pattern

4. Create throttle to protect resources:
   create_system_throttle(
     ruleset_name="Tactical",
     throttle_name="BI_ANALYTICS_LIMIT",
     throttle_type="DM",  # Delay Management
     limit=15,
     classification_criteria=[{
       "description": "BI Analytics workload throttle",
       "type": "APPL",
       "value": "BI_DASHBOARD",
       "operator": "I"
     }]
   )

5. Enable both rules:
   enable_filter_rule(ruleset_name="Tactical", filter_name="BI_DASHBOARD_FILTER")
   enable_throttle(ruleset_name="Tactical", throttle_name="BI_ANALYTICS_LIMIT")

6. Activate:
   activate_ruleset(ruleset_name="Tactical")

7. Verify:
   tdwm://ruleset/Tactical/filter/BI_DASHBOARD_FILTER
   → Confirm filter routes to BI_ANALYTICS workload

   tdwm://ruleset/Tactical/throttle/BI_ANALYTICS_LIMIT
   → Confirm throttle limit=15

8. Monitor (wait 10 minutes):
   show_tdwm_summary
   → BI queries now in BI_ANALYTICS workload (not DEFAULT)

   show_trottle_statistics
   → BI throttle active, limit enforced

Result: BI workload fully operational with routing and protection

Example 2: Emergency Workload Protection (Autonomous)

Scenario: Ad-hoc queries overloading system, need immediate throttling

Discovery: SKIP - Emergency situation requires fast action

Execution (Fast):

1. Use emergency workflow:
   tdwm://workflow/emergency-throttle

2. Create very restrictive throttle:
   create_system_throttle(
     ruleset_name="Tactical",
     throttle_name="EMERGENCY_ADHOC_LIMIT",
     throttle_type="DM",
     limit=3,  # Very restrictive
     classification_criteria=[{
       "description": "Emergency ad-hoc protection",
       "type": "USER",
       "value": "adhoc_*",  # All ad-hoc users
       "operator": "I"
     }]
   )

3. Enable and activate immediately:
   enable_throttle(ruleset_name="Tactical", throttle_name="EMERGENCY_ADHOC_LIMIT")
   activate_ruleset(ruleset_name="Tactical")

4. Monitor recovery (wait 60 seconds):
   show_physical_resources
   → CPU should drop below 90%

5. Gradual relaxation:
   modify_throttle_limit(
     ruleset_name="Tactical",
     throttle_name="EMERGENCY_ADHOC_LIMIT",
     new_limit=5
   )
   activate_ruleset(ruleset_name="Tactical")

   Wait, monitor, increase to 10 if stable

6. Cleanup when stable:
   disable_throttle(ruleset_name="Tactical", throttle_name="EMERGENCY_ADHOC_LIMIT")
   activate_ruleset(ruleset_name="Tactical")

Result: System stabilized in <5 minutes, gradual return to normal

Example 3: Seasonal Workload Activation (Autonomous)

Scenario: Year-end processing starting, activate YEAR_END workload with filters

Discovery:

1. list_WDs → YEAR_END workload exists (inactive)
2. tdwm://ruleset/Tactical/filter/YEAR_END_FILTER → Filter exists, disabled
3. tdwm://ruleset/Tactical/throttle/YEAR_END_LIMIT → Throttle exists, disabled
4. show_tdwm_summary → No year-end queries yet

Analysis:

  • Workload, filter, and throttle already exist (from last year)
  • Just need to enable them for the season

Execution (Autonomous):

1. Enable year-end filter:
   enable_filter_rule(
     ruleset_name="Tactical",
     filter_name="YEAR_END_FILTER"
   )

2. Enable year-end throttle:
   enable_throttle(
     ruleset_name="Tactical",
     throttle_name="YEAR_END_LIMIT"
   )

3. Activate:
   activate_ruleset(ruleset_name="Tactical")

4. Verify:
   tdwm://ruleset/Tactical/filter/YEAR_END_FILTER
   → Confirm enabled=true

   tdwm://ruleset/Tactical/throttle/YEAR_END_LIMIT
   → Confirm enabled=true

5. Monitor:
   show_tdwm_summary
   → Year-end queries routing to YEAR_END workload

6. Schedule deactivation (manual reminder):
   "Disable YEAR_END_FILTER and YEAR_END_LIMIT on January 5"

Result: Year-end workload active with proper routing and limits

Example 4: User-Based Workload Routing (Autonomous)

Scenario: Power users need dedicated workload with higher limits

Discovery:

1. list_active_WD → POWER_USERS workload exists
2. show_query_band → Power users DON'T set query band
3. tdwm://ruleset/Tactical/filters → No power user filter exists
4. Power users identified by account='power_user_acct'

Analysis:

  • Need to create filter using ACCT classification
  • Need throttle with higher limit (20 concurrent vs normal 10)

Execution (Autonomous):

1. Validate classification type:
   tdwm://reference/classification-types
   → Confirm "ACCT" is valid

2. Create filter for account-based routing:
   create_filter_rule(
     ruleset_name="Tactical",
     filter_name="POWER_USER_FILTER",
     classification_criteria=[{
       "description": "Power user account routing",
       "type": "ACCT",
       "value": "power_user_acct",
       "operator": "I"
     }],
     action_type="ACCEPT",
     workload_name="POWER_USERS"
   )

3. Create throttle with higher limit:
   create_system_throttle(
     ruleset_name="Tactical",
     throttle_name="POWER_USER_LIMIT",
     throttle_type="DM",
     limit=20,  # Higher than normal users
     classification_criteria=[{
       "description": "Power user concurrency limit",
       "type": "ACCT",
       "value": "power_user_acct",
       "operator": "I"
     }]
   )

4. Enable both:
   enable_filter_rule(ruleset_name="Tactical", filter_name="POWER_USER_FILTER")
   enable_throttle(ruleset_name="Tactical", throttle_name="POWER_USER_LIMIT")

5. Activate:
   activate_ruleset(ruleset_name="Tactical")

6. Verify:
   tdwm://ruleset/Tactical/filter/POWER_USER_FILTER
   → Confirm routes to POWER_USERS workload

   tdwm://ruleset/Tactical/throttle/POWER_USER_LIMIT
   → Confirm limit=20

7. Monitor:
   show_tdwm_summary
   → Power user queries in POWER_USERS workload

Result: Power users have dedicated workload with higher concurrency

Example 5: Maintenance Window Workload (Autonomous)

Scenario: During Saturday maintenance, only maintenance queries should run

Discovery:

1. list_WDs → MAINTENANCE_ONLY workload exists
2. tdwm://ruleset/Tactical/filters → No maintenance filter exists
3. Need to block all non-maintenance queries during window

Analysis:

  • Need filter to route maintenance user queries to MAINTENANCE_ONLY
  • Need filter to REJECT all other queries
  • Will enable for maintenance window, disable after

Execution (Autonomous):

1. Create maintenance filter (ACCEPT):
   create_filter_rule(
     ruleset_name="Tactical",
     filter_name="MAINTENANCE_ACCEPT",
     classification_criteria=[{
       "description": "Allow maintenance user",
       "type": "USER",
       "value": "maint_user",
       "operator": "I"
     }],
     action_type="ACCEPT",
     workload_name="MAINTENANCE_ONLY"
   )

2. Create blocking filter (REJECT):
   create_filter_rule(
     ruleset_name="Tactical",
     filter_name="MAINTENANCE_BLOCK_ALL",
     classification_criteria=[{
       "description": "Block all non-maintenance queries",
       "type": "USER",
       "value": "*",  # All users
       "operator": "I"
     }],
     action_type="REJECT",
     workload_name=None  # REJECT doesn't need workload
   )

3. Keep both DISABLED initially (will enable during maintenance):
   # Don't enable yet - wait for maintenance window

4. Activate (makes filters available, but disabled):
   activate_ruleset(ruleset_name="Tactical")

5. At maintenance start (Saturday 2 AM):
   enable_filter_rule(ruleset_name="Tactical", filter_name="MAINTENANCE_ACCEPT")
   enable_filter_rule(ruleset_name="Tactical", filter_name="MAINTENANCE_BLOCK_ALL")
   activate_ruleset(ruleset_name="Tactical")

6. At maintenance end (Saturday 10 AM):
   disable_filter_rule(ruleset_name="Tactical", filter_name="MAINTENANCE_ACCEPT")
   disable_filter_rule(ruleset_name="Tactical", filter_name="MAINTENANCE_BLOCK_ALL")
   activate_ruleset(ruleset_name="Tactical")

7. Verify normal operations resumed:
   show_tdwm_summary
   → All workloads operational

Result: Maintenance window enforced, normal operations restored after

Best Practices

General Principles

  • Workload definitions created by Teradata TDWM, but filters/throttles make them functional
  • Always create both filter (routing) and throttle (protection) for new workloads
  • Test workload implementation in non-production first
  • Document business justification for each workload
  • Name filters and throttles clearly (e.g., "BI_DASHBOARD_FILTER", "BI_ANALYTICS_LIMIT")

Discovery Before Execution (NEW ✨)

  • ALWAYS check existing filters and throttles before creating new ones
  • Use tdwm://ruleset/{name}/filters and tdwm://ruleset/{name}/throttles to explore
  • Check tdwm://ruleset/{name}/pending-changes before activating
  • Understand current workload distribution before implementing changes

Template Usage (NEW ✨)

  • Use templates for common patterns (application-based, user-based)
  • Templates reduce errors by 50% compared to custom creation
  • Customize template parameters for your specific workload needs

Execution Safety (NEW ✨)

  • Validate classification types against tdwm://reference/classification-types
  • Validate operators against tdwm://reference/operators
  • Start with conservative throttle limits, increase gradually
  • Test filter routing with sample queries before full activation
  • Keep emergency rollback plan (disable filter + activate)

Validation (NEW ✨)

  • Always read back filter/throttle after creation
  • Confirm enabled state matches intention
  • Monitor workload distribution for 15-30 minutes after activation
  • Verify queries routing as expected using show_tdwm_summary

Monitoring

  • Monitor workload effectiveness using show_tasm_statistics
  • Check throttle impact using show_trottle_statistics
  • Review classification accuracy using show_tasm_even_history
  • Set alerts for workload performance SLA violations
  • Coordinate workload changes with application teams
  • Document workload lifecycle events

Related Skills

  • Use optimize-throttles skill for adjusting throttle limits on existing workloads
  • Use tune-workloads skill for adding classification criteria to existing filters
  • Use emergency-response skill for crisis situations
  • Use discover-configuration skill to inventory existing workload configurations
  • Use configure-filters skill for detailed filter creation guidance
  • Use configure-throttles skill for detailed throttle creation guidance