Scalable time-series analytics with InfluxDB

Tutorial overview

Stream and analyze pool OHLCV in InfluxDB with hourly/minute intervals and live Grafana dashboards.

From real-time monitoring to historical analysis

While InfluxDB is a champion of real-time data, its power extends far beyond live dashboards. It provides a highly optimized engine for storing and querying massive time-series datasets, making it a perfect middle-ground between the local analytics of DuckDB and the enterprise scale of ClickHouse.

Looking for other analytics solutions? Check out our full list of API Tutorials for more step-by-step guides.

This tutorial will guide you through building a production-grade ETL pipeline to populate InfluxDB with a substantial historical dataset, enabling both high-performance queries and real-time visualization. The goal: By the end of this guide, you will have a scalable analytics pipeline that can:

Ingest 1-hour OHLCV data for the top 500 Uniswap v3 pools over a 14-day period.
Run complex time-series analysis using Python and the Flux language.
Visualize the data in a live-updating Grafana dashboard.

Step 1: Docker Setup

Get InfluxDB and Grafana running in seconds with Docker.

Step 2: ETL Pipeline

Create a robust data pipeline for ingesting large time-series datasets.

Step 3: Programmatic Analysis

Analyze your time-series data with InfluxDB’s Python client.

Step 4: Live Dashboard

Build a real-time dashboard to monitor crypto pools.

Step 1: Setting up InfluxDB and Grafana

We’ll use Docker Compose to spin up both services. First, create a new directory named INFLUXDB in your project root. Inside that directory, create a file named docker-compose.yml with the following content.

INFLUXDB/docker-compose.yml

services:
  influxdb:
    image: influxdb:2.7
    container_name: influxdb
    ports:
      - "8087:8086"
    volumes:
      - influxdb-data:/var/lib/influxdb2
    environment:
      - DOCKER_INFLUXDB_INIT_MODE=setup
      - DOCKER_INFLUXDB_INIT_USERNAME=my-user
      - DOCKER_INFLUXDB_INIT_PASSWORD=my-password
      - DOCKER_INFLUXDB_INIT_ORG=my-org
      - DOCKER_INFLUXDB_INIT_BUCKET=crypto-data
      - DOCKER_INFLUXDB_INIT_ADMIN_TOKEN=my-super-secret-token

  grafana:
    image: grafana/grafana:latest
    container_name: grafana
    ports:
      - "3000:3000"
    volumes:
      - grafana-data:/var/lib/grafana

volumes:
  influxdb-data:
  grafana-data:

Run the following command from the root of the project to start the containers:

docker-compose -f INFLUXDB/docker-compose.yml up -d

Once running, you can access:

InfluxDB UI: http://localhost:8087
Grafana UI: http://localhost:3000 (login with admin/admin)

Use the token my-super-secret-token to connect to InfluxDB.

We use port 8087 for InfluxDB to avoid potential conflicts with other services that might be using the default port 8086.

Step 2: Build the Python ETL pipeline

Create a new file named INFLUXDB/build_influxdb_db.py. This script is built to be robust and efficient, capable of ingesting large volumes of time-series data from the DexPaprika API into your InfluxDB instance. It uses two endpoints: pool search with a dex_name filter to discover pools, and the Pool OHLCV Data endpoint to fetch historical price data.

This script needs a free API key. Without one, OHLCV reaches back only 24 hours at 1h and up, and the script asks for six days of hourly candles. Export the key before running it; it goes in the Authorization header on its own. On Pro, point API_BASE_URL at https://api-pro.dexpaprika.com. Limits per plan: OHLCV limits by plan.

export DEXPAPRIKA_API_KEY="your-key"

INFLUXDB/build_influxdb_db.py

import influxdb_client
from influxdb_client.client.write_api import SYNCHRONOUS
import requests
from datetime import datetime, timedelta, timezone
import time
import logging
import asyncio
import aiohttp
from typing import List, Dict
import math
import os

# --- Configuration ---
API_BASE_URL = "https://api.dexpaprika.com"
API_KEY = os.environ["DEXPAPRIKA_API_KEY"]  # free key from console.dexpaprika.com
NETWORK = "ethereum"
DEX_ID = "uniswap_v3"
HISTORY_DAYS = 6        # a free key reaches 7 days back; stay inside it
TOP_POOLS_LIMIT = 500   # Focus on top 500 pools by volume
BATCH_SIZE = 15         # Process pools in smaller batches
CONCURRENT_REQUESTS = 3 # Concurrent requests for API calls
OHLCV_API_LIMIT = 100   # API limit for OHLCV requests
INTERVAL = "1h"         # 1-hour intervals

# InfluxDB Configuration
INFLUX_URL = "http://localhost:8087"
INFLUX_TOKEN = "my-super-secret-token"
INFLUX_ORG = "my-org"
INFLUX_BUCKET = "crypto-data"

# Setup logging
logging.basicConfig(level=logging.INFO, format='%(asctime)s - %(levelname)s - %(message)s')

async def fetch_with_retry(session: aiohttp.ClientSession, url: str, params: Dict = None, retries=5, backoff_factor=1.0):
    """Generic async fetch function with exponential backoff."""
    for attempt in range(retries):
        try:
            async with session.get(url, params=params, timeout=30) as response:
                response.raise_for_status()
                return await response.json()
        except (aiohttp.ClientError, asyncio.TimeoutError) as e:
            if attempt == retries - 1:
                logging.error(f"Final attempt failed for {url}: {e}")
                raise
            sleep_time = backoff_factor * (2 ** attempt)
            logging.warning(f"Request to {url} failed: {e}. Retrying in {sleep_time:.2f}s...")
            await asyncio.sleep(sleep_time)
    return {}

class InfluxDBETL:
    def __init__(self):
        self.client = influxdb_client.InfluxDBClient(url=INFLUX_URL, token=INFLUX_TOKEN, org=INFLUX_ORG)
        self.write_api = self.client.write_api(write_options=SYNCHRONOUS)
        self.api_semaphore = asyncio.Semaphore(CONCURRENT_REQUESTS)
        self._ensure_bucket_exists()

    def _ensure_bucket_exists(self):
        """Checks if the bucket exists and creates it if not."""
        logging.info(f"Ensuring bucket '{INFLUX_BUCKET}' exists...")
        buckets_api = self.client.buckets_api()
        bucket = buckets_api.find_bucket_by_name(INFLUX_BUCKET)
        if not bucket:
            logging.warning(f"Bucket '{INFLUX_BUCKET}' not found. Creating it...")
            buckets_api.create_bucket(bucket_name=INFLUX_BUCKET, org=INFLUX_ORG)
            logging.info(f"Bucket '{INFLUX_BUCKET}' created successfully.")
        else:
            logging.info(f"Bucket '{INFLUX_BUCKET}' already exists.")

    def clear_bucket_data(self):
        """Deletes all data from the 'ohlcv' measurement in the bucket."""
        logging.info(f"Clearing existing data from measurement 'ohlcv' in bucket '{INFLUX_BUCKET}'...")
        try:
            delete_api = self.client.delete_api()
            start = "1970-01-01T00:00:00Z"
            stop = datetime.now(timezone.utc).strftime('%Y-%m-%dT%H:%M:%SZ')
            delete_api.delete(start, stop, '_measurement="ohlcv"', bucket=INFLUX_BUCKET, org=INFLUX_ORG)
            logging.info("Existing data cleared successfully.")
        except Exception as e:
            logging.error(f"Could not clear data from bucket: {e}")

    async def fetch_top_pools(self) -> List[Dict]:
        """Fetch top pools by volume from the specified DEX, handling cursor pagination."""
        logging.info(f"Fetching top {TOP_POOLS_LIMIT} pools for {DEX_ID} on {NETWORK}...")
        all_pools = []
        cursor = None
        async with aiohttp.ClientSession(headers={"Authorization": API_KEY}) as session:
            while len(all_pools) < TOP_POOLS_LIMIT:
                # The DEX is a filter now: /networks/{network}/dexes/{dex}/pools was
                # removed and returns 410. detailed=true keeps name/symbol on tokens[].
                url = f"{API_BASE_URL}/networks/{NETWORK}/pools/search"
                params = {"dex_name": DEX_ID, "limit": 100, "order_by": "volume_usd_24h",
                          "sort": "desc", "detailed": "true"}
                if cursor:
                    params["cursor"] = cursor
                try:
                    data = await fetch_with_retry(session, url, params=params)
                    pools = data.get('results', [])
                    if not pools:
                        break
                    all_pools.extend(pools)
                    logging.info(f"Fetched a page, got {len(pools)} pools. Total: {len(all_pools)}")
                    if len(all_pools) >= TOP_POOLS_LIMIT:
                        all_pools = all_pools[:TOP_POOLS_LIMIT]
                        break
                    if not data.get('has_next_page'):
                        break
                    cursor = data.get('next_cursor')
                    await asyncio.sleep(0.5)  # Be respectful to the API
                except Exception as e:
                    logging.error(f"Error fetching a page of pools: {e}")
                    break
        logging.info(f"Finished fetching pools. Total: {len(all_pools)}")
        return all_pools

    async def fetch_pool_ohlcv_paginated(self, session: aiohttp.ClientSession, pool_address: str) -> List[Dict]:
        """Fetch complete OHLCV data for a pool using intelligent, dynamic windowing."""
        async with self.api_semaphore:
            final_end_time = datetime.now(timezone.utc)
            current_start_time = final_end_time - timedelta(days=HISTORY_DAYS)
            all_ohlcv = []

            try:
                if 'h' in INTERVAL:
                    interval_value = int(INTERVAL.replace('h', ''))
                    time_delta_per_call = timedelta(hours=OHLCV_API_LIMIT * interval_value)
                elif 'm' in INTERVAL:
                    interval_value = int(INTERVAL.replace('m', ''))
                    time_delta_per_call = timedelta(minutes=OHLCV_API_LIMIT * interval_value)
                else:
                    raise ValueError(f"Unsupported interval format: {INTERVAL}")
            except ValueError as e:
                logging.error(f"Invalid INTERVAL format: {e}. Defaulting to 1 hour.")
                time_delta_per_call = timedelta(hours=OHLCV_API_LIMIT * 1)

            total_expected_calls = math.ceil((final_end_time - current_start_time) / time_delta_per_call) if time_delta_per_call.total_seconds() > 0 else 0
            call_count = 0
            while current_start_time < final_end_time:
                call_count += 1
                batch_end_time = min(current_start_time + time_delta_per_call, final_end_time)
                
                logging.info(f"  [Pool {pool_address}] Fetching window {call_count}/{total_expected_calls}: {current_start_time.date()} to {batch_end_time.date()}")

                url = f"{API_BASE_URL}/networks/{NETWORK}/pools/{pool_address}/ohlcv"
                params = {
                    "start": current_start_time.strftime('%Y-%m-%dT%H:%M:%SZ'),
                    "end": batch_end_time.strftime('%Y-%m-%dT%H:%M:%SZ'),
                    "interval": INTERVAL,
                    "limit": OHLCV_API_LIMIT
                }
                try:
                    batch_data = await fetch_with_retry(session, url, params=params)
                    if batch_data:
                        for record in batch_data:
                            record['network'] = NETWORK
                            record['pool_address'] = pool_address
                        all_ohlcv.extend(batch_data)
                except Exception as e:
                    logging.warning(f"Could not fetch OHLCV batch for {pool_address}: {e}")
                
                current_start_time = batch_end_time
                await asyncio.sleep(0.75)  # Crucial delay to prevent rate-limiting
            
            logging.info(f"Pool {pool_address}: collected {len(all_ohlcv)} OHLCV records.")
            return all_ohlcv

    async def fetch_pool_ohlcv_batch(self, pool_addresses: List[str]) -> List[Dict]:
        """Fetch OHLCV data for multiple pools concurrently."""
        logging.info(f"Fetching {INTERVAL} OHLCV for {len(pool_addresses)} pools...")
        all_ohlcv = []
        async with aiohttp.ClientSession(headers={"Authorization": API_KEY}) as session:
            tasks = [self.fetch_pool_ohlcv_paginated(session, addr) for addr in pool_addresses]
            results = await asyncio.gather(*tasks, return_exceptions=True)
            for i, result in enumerate(results):
                if isinstance(result, list):
                    all_ohlcv.extend(result)
                elif isinstance(result, Exception):
                    logging.warning(f"OHLCV fetch failed for pool {pool_addresses[i]}: {result}")
        return all_ohlcv

    def load_ohlcv_data(self, ohlcv_data: List[Dict], pools_map: Dict):
        """Load OHLCV data into InfluxDB."""
        if not ohlcv_data:
            logging.warning("No OHLCV data to load.")
            return

        points = []
        for record in ohlcv_data:
            pool_id = record.get('pool_address')
            pair = pools_map.get(pool_id, "Unknown/Unknown")
            
            point = (
                influxdb_client.Point("ohlcv")
                .tag("pool_id", pool_id)
                .tag("pair", pair)
                .field("open", float(record['open']))
                .field("high", float(record['high']))
                .field("low", float(record['low']))
                .field("close", float(record['close']))
                .field("volume", float(record.get('volume', 0)))
                .time(record['time_close'])
            )
            points.append(point)
        
        if points:
            self.write_api.write(bucket=INFLUX_BUCKET, org=INFLUX_ORG, record=points)
            logging.info(f"Wrote {len(points)} data points to InfluxDB.")

    async def run_etl(self):
        """Run the complete ETL process."""
        self.clear_bucket_data()
        logging.info(f"Starting InfluxDB ETL process for top {TOP_POOLS_LIMIT} pools...")
        pools = await self.fetch_top_pools()
        if pools:
            pools_map = {
                pool['id']: f"{pool['tokens'][0]['symbol']}/{pool['tokens'][1]['symbol']}"
                for pool in pools if len(pool.get('tokens', [])) >= 2
            }
            pool_addresses = [pool['id'] for pool in pools if pool.get('id')]
            
            for i in range(0, len(pool_addresses), BATCH_SIZE):
                batch_addresses = pool_addresses[i:i + BATCH_SIZE]
                batch_num = (i // BATCH_SIZE) + 1
                total_batches = (len(pool_addresses) + BATCH_SIZE - 1) // BATCH_SIZE
                logging.info(f"Processing OHLCV batch {batch_num}/{total_batches} ({len(batch_addresses)} pools)")
                
                ohlcv_data = await self.fetch_pool_ohlcv_batch(batch_addresses)
                self.load_ohlcv_data(ohlcv_data, pools_map)
                
                if i + BATCH_SIZE < len(pool_addresses):
                    logging.info(f"--- Finished batch {batch_num}, sleeping for 10 seconds ---")
                    await asyncio.sleep(10)
        
        logging.info("ETL process completed!")

async def main():
    etl = InfluxDBETL()
    await etl.run_etl()

if __name__ == "__main__":
    # pip install influxdb-client aiohttp requests pandas
    asyncio.run(main())

This script is built for performance and reliability, using several best practices common in data pipelines:

Asynchronous operations: By using asyncio and aiohttp, the script can make many API requests concurrently instead of one by one.
Dynamic windowing: The fetch_pool_ohlcv_paginated function calculates how much data to request per API call to ensure complete history is fetched efficiently.
Concurrency control & throttling: An asyncio.Semaphore, combined with carefully tuned BATCH_SIZE and asyncio.sleep() calls, keeps the script under the free tier’s 15 requests a minute keyless, or 30 with a free key. Pro raises that to 500 a minute, which lets you shorten the sleeps and finish the backfill sooner.
Resiliency: The fetch_with_retry function automatically retries failed requests with an exponential backoff delay.
Data Integrity: The script automatically clears old data from the bucket before each run to ensure a clean, consistent dataset.

Set up the Python environment

Before running the ETL script, it’s a critical best practice to create an isolated Python environment to manage dependencies.

Create a virtual environment: Open your terminal in the project root and run:
python3 -m venv venv
Activate the environment:
- On macOS and Linux:
  source venv/bin/activate
- On Windows:
  .\venv\Scripts\activate
Your terminal prompt should now be prefixed with (venv), indicating that the virtual environment is active.
Install required libraries: Now, install the necessary Python packages inside the activated environment:
pip install influxdb-client aiohttp requests pandas

Run the script to start streaming data into your InfluxDB instance. This may take several minutes.

python INFLUXDB/build_influxdb_db.py

Step 3: Programmatic analysis with Python

While the InfluxDB UI is great for exploration, the real power comes from programmatic access. The influxdb-client library for Python allows you to run complex Flux queries and integrate the data into other tools or scripts. We’ve created a query_influxdb.py script to demonstrate how to connect to your database and perform analysis on the hourly data.

INFLUXDB/query_influxdb.py

import influxdb_client
import pandas as pd

# --- InfluxDB Configuration ---
INFLUX_URL = "http://localhost:8087"
INFLUX_TOKEN = "my-super-secret-token"
INFLUX_ORG = "my-org"
INFLUX_BUCKET = "crypto-data"

def run_flux_query(client: influxdb_client.InfluxDBClient, query: str):
    """Helper function to execute a Flux query and return a pandas DataFrame."""
    try:
        query_api = client.query_api()
        result = query_api.query_data_frame(query, org=INFLUX_ORG)
        if isinstance(result, list): # Handle multiple dataframes in result
            return pd.concat(result, ignore_index=True) if result else pd.DataFrame()
        return result
    except Exception as e:
        print(f"Error running query: {e}")
        return pd.DataFrame()

def main():
    """Connects to InfluxDB and runs sample analytics queries."""
    client = influxdb_client.InfluxDBClient(url=INFLUX_URL, token=INFLUX_TOKEN, org=INFLUX_ORG)
    
    print("=== InfluxDB Python Analytics Demo ===\n")

    # --- Query 1: Find available trading pairs ---
    print("--- 1. Finding available trading pairs ---")
    list_pairs_query = f'''
    import "influxdata/influxdb/schema"

    schema.tagValues(
      bucket: "{INFLUX_BUCKET}",
      tag: "pair",
      start: -14d
    )
    '''
    pairs_df = run_flux_query(client, list_pairs_query)
    
    if not pairs_df.empty:
        available_pairs = pairs_df['_value'].tolist()
        print(f"Found {len(available_pairs)} pairs. Examples: {available_pairs[:5]}")
        # Use the first available pair for subsequent queries
        target_pair = available_pairs[0]
    else:
        print("No pairs found. Please run the ingestion script first.")
        print("Using 'WETH/USDC' as a placeholder for query examples.")
        target_pair = "WETH/USDC" # Fallback for demo
    
    print(f"\n--- Using pair '{target_pair}' for next queries ---\n")

    # --- Query 2: Get raw data for the target pool ---
    print(f"--- 2. Raw OHLCV data for {target_pair} ---")
    raw_data_query = f'''
    from(bucket: "{INFLUX_BUCKET}")
      |> range(start: -3d) // Limit to last 3 days for brevity
      |> filter(fn: (r) => r._measurement == "ohlcv")
      |> filter(fn: (r) => r.pair == "{target_pair}")
      |> pivot(rowKey:["_time"], columnKey: ["_field"], valueColumn: "_value")
      |> sort(columns: ["_time"], desc: true)
      |> limit(n: 10)
    '''
    raw_df = run_flux_query(client, raw_data_query)
    print("Last 10 records:")
    if not raw_df.empty and all(c in raw_df.columns for c in ['_time', 'open', 'high', 'low', 'close', 'volume']):
        print(raw_df[['_time', 'open', 'high', 'low', 'close', 'volume']])
    else:
        print("Could not retrieve raw data. Please check if the ingestion was successful.")
    print("\n")

    # --- Query 3: Calculate 12-hour moving average ---
    print(f"--- 3. 12-Hour moving average for {target_pair} close price ---")
    moving_avg_query = f'''
    from(bucket: "{INFLUX_BUCKET}")
      |> range(start: -14d)
      |> filter(fn: (r) => r._measurement == "ohlcv" and r._field == "close" and r.pair == "{target_pair}")
      |> timedMovingAverage(every: 1h, period: 12h)
      |> sort(columns: ["_time"], desc: true)
      |> limit(n: 10)
    '''
    ma_df = run_flux_query(client, moving_avg_query)
    print("Last 10 moving average values:")
    if not ma_df.empty and all(c in ma_df.columns for c in ['_time', '_value']):
        print(ma_df[['_time', '_value']])
    else:
        print("Could not retrieve moving average data.")


if __name__ == "__main__":
    # Ensure you have the required libraries:
    # pip install influxdb-client pandas
    main()

Run the script to see how you can query your data with Python:

python INFLUXDB/query_influxdb.py

Step 4: Visualizing data in Grafana

Open Grafana at http://localhost:3000.
Go to Connections > Add new connection > InfluxDB.
Configure the connection:
- Name: InfluxDB_Crypto
- Query Language: Flux
- URL: http://influxdb:8086 (use the Docker service name)
- Under Auth, toggle Basic auth off.
- In the Custom HTTP Headers section, add a header:
  - Header: Authorization
  - Value: Token my-super-secret-token
- Enter your InfluxDB Organization (my-org) and default Bucket (crypto-data).
Click Save & Test. You should see a “Bucket found” confirmation.
Now, let’s create a dashboard. In the left-hand menu, click the + icon and select Dashboard.
Click on the Add new panel button.
In the new panel view, ensure your InfluxDB_Crypto data source is selected at the top.
Below the graph, you’ll see a query editor. You can switch to the Script editor by clicking the pencil icon on the right.
Paste the following query into the editor. This query will plot the raw closing price for the AAVE/USDC trading pair.
from(bucket: "crypto-data") |> range(start: v.timeRangeStart, stop: v.timeRangeStop) |> filter(fn: (r) => r._measurement == "ohlcv" and r._field == "close" and r.pair == "AAVE/USDC") |> sort(columns: ["_time"], desc: false)
Troubleshooting Tip: If you initially see “No data,” there are two common reasons:
1. Time Range: Ensure the time picker at the top right is set to “Last 7 days” or wider, not a shorter period like “Last 6 hours.”
2. Trading Pair: The default WETH/USDC pair used in the original tutorial may not have been in the top 500 pools fetched by the script. The query above uses AAVE/USDC, which is more likely to be present. You can find other available pairs by running the query_influxdb.py script.
At the top right of the dashboard page, set the time range to Last 7 days to ensure you see all the historical data you ingested.
You should now see the data appear in the panel. You can give the panel a title (e.g., “AAVE/USDC Close Price”) in the settings on the right.
Click Apply to save the panel to your dashboard. You can now add more panels for other queries.

What you’ve built

You now have a powerful, scalable analytics pipeline for time-series crypto data. You’ve combined a production-grade Python ETL script with the industry-standard tools for time-series data storage (InfluxDB) and visualization (Grafana). Key achievements:

Built a production-ready ETL pipeline: You have a reusable, high-performance Python script that can populate a time-series database from any supported DEX.
Unlocked programmatic time-series analysis: You can now perform complex analytical queries on large time-series datasets using Python and Flux.
Mastered a scalable analytics workflow: This pipeline provides a solid foundation for building live dashboards, conducting in-depth market research, and developing sophisticated monitoring or trading algorithms.
Enabled live data visualization: You’ve connected your database to Grafana, the leading open-source tool for observability and data visualization.

FAQs

Why choose InfluxDB over ClickHouse for this use case?

InfluxDB excels at time-series ingestion and dashboards; it’s ideal for hourly/minute data with Grafana visualization.

How much history should I store?

Keep recent weeks/months for dashboards and archive older data elsewhere; adjust by bucket retention.

What interval is best for the ETL?

1h is a sensible default; lower to 15m/5m for more granularity if needed.

How do I keep the dashboard fresh?

Schedule the ETL to append new windows, and set Grafana panels to auto-refresh.

Get Started

Knowledge Base

Tutorials

SDKs

Tools

Company

Scalable time-series analytics with InfluxDB

Tutorial overview

From real-time monitoring to historical analysis

Step 1: Docker Setup

Step 2: ETL Pipeline

Step 3: Programmatic Analysis

Step 4: Live Dashboard

Step 1: Setting up InfluxDB and Grafana

Step 2: Build the Python ETL pipeline

Set up the Python environment

Step 3: Programmatic analysis with Python

Step 4: Visualizing data in Grafana

What you’ve built

FAQs

​Tutorial overview

​From real-time monitoring to historical analysis

Step 1: Docker Setup

Step 2: ETL Pipeline

Step 3: Programmatic Analysis

Step 4: Live Dashboard

​Step 1: Setting up InfluxDB and Grafana

​Step 2: Build the Python ETL pipeline

​Set up the Python environment

​Step 3: Programmatic analysis with Python

​Step 4: Visualizing data in Grafana

​What you’ve built

​FAQs

Tutorial overview

From real-time monitoring to historical analysis

Step 1: Setting up InfluxDB and Grafana

Step 2: Build the Python ETL pipeline

Set up the Python environment

Step 3: Programmatic analysis with Python

Step 4: Visualizing data in Grafana

What you’ve built

FAQs