File size: 7,517 Bytes

fd357f4

-- Cassandra Schema for DTO Durable Storage
-- Data Transfer Operations run history and metrics

-- Keyspace for DTO operations
CREATE KEYSPACE IF NOT EXISTS dto_operations 
WITH replication = {
    'class': 'SimpleStrategy',
    'replication_factor': 2
}
AND durable_writes = true;

-- Table for run metadata and status
CREATE TABLE IF NOT EXISTS dto_operations.runs (
    run_id text,
    job_id text,
    manifest_path text,
    data_class text,          -- CLASS_A, CLASS_B, CLASS_C
    environment text,         -- staging, prod, archive
    status text,              -- planned, in_progress, completed, failed, rolled_back
    start_time timestamp,
    end_time timestamp,
    initiated_by text,
    approvers set<text>,
    data_size_bytes bigint,
    estimated_duration text,
    final_status text,
    total_duration_seconds int,
    average_throughput_mbps double,
    artifacts list<text>,
    metadata map<text, text>,
    PRIMARY KEY ((run_id), start_time)
) WITH CLUSTERING ORDER BY (start_time DESC)
   AND default_time_to_live = 2592000; -- 30 days TTL

-- Table for performance metrics (time-series data)
CREATE TABLE IF NOT EXISTS dto_operations.metrics (
    run_id text,
    metric_name text,         -- throughput_mbps, checksum_time_seconds, validation_passed
    bucket timestamp,         -- minute precision bucket
    timestamp timestamp,
    value double,
    labels map<text, text>,   -- source_host, target_host, transfer_method
    PRIMARY KEY ((run_id, metric_name), bucket, timestamp)
) WITH CLUSTERING ORDER BY (bucket DESC, timestamp DESC)
   AND default_time_to_live = 604800; -- 7 days TTL

-- Table for event history (durable event store)
CREATE TABLE IF NOT EXISTS dto_operations.events (
    event_id text,
    event_type text,          -- RUN_PLANNED, PREFLIGHT_PASSED, etc.
    timestamp timestamp,
    run_id text,
    job_id text,
    payload text,             -- JSON payload
    PRIMARY KEY ((run_id), timestamp, event_type)
) WITH CLUSTERING ORDER BY (timestamp DESC, event_type ASC)
   AND default_time_to_live = 2592000; -- 30 days TTL

-- Table for SLI (Service Level Indicator) tracking
CREATE TABLE IF NOT EXISTS dto_operations.slis (
    sli_name text,            -- throughput_mbps, validation_success_rate
    run_id text,
    timestamp timestamp,
    expected_value double,
    actual_value double,
    breach_duration_seconds int,
    PRIMARY KEY ((sli_name), run_id, timestamp)
) WITH CLUSTERING ORDER BY (run_id DESC, timestamp DESC)
   AND default_time_to_live = 2592000; -- 30 days TTL

-- Table for artifact references
CREATE TABLE IF NOT EXISTS dto_operations.artifacts (
    run_id text,
    artifact_type text,       -- logs, reports, snapshots, checksums
    artifact_path text,
    created_at timestamp,
    size_bytes bigint,
    checksum text,
    PRIMARY KEY ((run_id, artifact_type), created_at)
) WITH CLUSTERING ORDER BY (created_at DESC)
   AND default_time_to_live = 2592000; -- 30 days TTL

-- Table for job queue state (durable job tracking)
CREATE TABLE IF NOT EXISTS dto_operations.jobs (
    job_id text,
    status text,              -- pending, in_progress, completed, failed
    run_id text,
    created_at timestamp,
    updated_at timestamp,
    priority int,             -- 1-10 priority
    retry_count int,
    error_message text,
    next_retry_time timestamp,
    PRIMARY KEY ((job_id), created_at)
) WITH CLUSTERING ORDER BY (created_at DESC)
   AND default_time_to_live = 604800; -- 7 days TTL

-- Materialized views for common queries

-- View for finding runs by status
CREATE MATERIALIZED VIEW IF NOT EXISTS dto_operations.runs_by_status AS
    SELECT run_id, job_id, status, start_time, environment, data_class
    FROM dto_operations.runs
    WHERE status IS NOT NULL AND run_id IS NOT NULL AND start_time IS NOT NULL
    PRIMARY KEY ((status), run_id, start_time)
    WITH CLUSTERING ORDER BY (run_id DESC, start_time DESC);

-- View for finding runs by environment
CREATE MATERIALIZED VIEW IF NOT EXISTS dto_operations.runs_by_environment AS
    SELECT run_id, job_id, status, start_time, environment, data_class
    FROM dto_operations.runs
    WHERE environment IS NOT NULL AND run_id IS NOT NULL AND start_time IS NOT NULL
    PRIMARY KEY ((environment), run_id, start_time)
    WITH CLUSTERING ORDER BY (run_id DESC, start_time DESC);

-- View for finding runs by data class
CREATE MATERIALIZED VIEW IF NOT EXISTS dto_operations.runs_by_data_class AS
    SELECT run_id, job_id, status, start_time, environment, data_class
    FROM dto_operations.runs
    WHERE data_class IS NOT NULL AND run_id IS NOT NULL AND start_time IS NOT NULL
    PRIMARY KEY ((data_class), run_id, start_time)
    WITH CLUSTERING ORDER BY (run_id DESC, start_time DESC);

-- View for recent runs across all environments
CREATE MATERIALIZED VIEW IF NOT EXISTS dto_operations.recent_runs AS
    SELECT run_id, job_id, status, start_time, environment, data_class
    FROM dto_operations.runs
    WHERE run_id IS NOT NULL AND start_time IS NOT NULL
    PRIMARY KEY ((environment, data_class), start_time, run_id)
    WITH CLUSTERING ORDER BY (start_time DESC);

-- Indexes for better query performance
CREATE INDEX IF NOT EXISTS ON dto_operations.runs (job_id);
CREATE INDEX IF NOT EXISTS ON dto_operations.runs (initiated_by);
CREATE INDEX IF NOT EXISTS ON dto_operations.events (event_type);
CREATE INDEX IF NOT EXISTS ON dto_operations.events (job_id);
CREATE INDEX IF NOT EXISTS ON dto_operations.metrics (metric_name);
CREATE INDEX IF NOT EXISTS ON dto_operations.artifacts (artifact_type);

-- User-defined types for complex data
CREATE TYPE IF NOT EXISTS dto_operations.throughput_sample (
    timestamp timestamp,
    value double,
    window_seconds int
);

CREATE TYPE IF NOT EXISTS dto_operations.checksum_result (
    algorithm text,
    expected text,
    actual text,
    matched boolean,
    verified_at timestamp
);

-- Functions for common operations
CREATE FUNCTION IF NOT EXISTS dto_operations.avg_throughput(throughputs list<double>)
    RETURNS NULL ON NULL INPUT
    RETURNS double
    LANGUAGE java
    AS 'if (throughputs == null || throughputs.isEmpty()) return 0.0; 
        double sum = 0.0; 
        for (double t : throughputs) sum += t; 
        return sum / throughputs.size();';

CREATE FUNCTION IF NOT EXISTS dto_operations.percent_complete(transferred bigint, total bigint)
    RETURNS NULL ON NULL INPUT
    RETURNS double
    LANGUAGE java
    AS 'if (total == 0) return 0.0; 
        return (double) transferred / total * 100.0;';

-- Insert sample data for testing
INSERT INTO dto_operations.runs (
    run_id, job_id, manifest_path, data_class, environment, status,
    start_time, initiated_by, data_size_bytes, estimated_duration
) VALUES (
    'test-run-001', 'test-job-001', '/manifests/class_a/test.yaml', 'CLASS_A', 'staging', 'completed',
    toTimestamp(now()), 'prometheus', 107374182400, '2h'
);

INSERT INTO dto_operations.metrics (
    run_id, metric_name, bucket, timestamp, value, labels
) VALUES (
    'test-run-001', 'throughput_mbps', 
    dateOf(now()) - 1 * 60 * 1000, -- 1 minute bucket
    toTimestamp(now()), 604.0,
    {'source_host': 'vast2', 'target_host': 'vast1', 'transfer_method': 'ssh+dd'}
);

-- Grant permissions to DTO service account
CREATE ROLE IF NOT EXISTS dto_service WITH LOGIN = true AND PASSWORD = '${CASSANDRA_DTO_PASSWORD}';
GRANT ALL PERMISSIONS ON KEYSPACE dto_operations TO dto_service;
GRANT EXECUTE ON ALL FUNCTIONS IN KEYSPACE dto_operations TO dto_service;